[12:13] <pitti> mathiaz: I've seen it before, and it makes sense to me at least
[12:13] <pitti> mathiaz: I'll sync it right now then, just in case
[12:14] <Riddell> pitti: is the pkgmaintainermangler error fixed?
[12:14] <pitti> Riddell: yep
[12:14] <pitti> Riddell: I gave back and rebuilt the important stuff for CDs
[12:14] <mjg59> seb128: There's something up with epiphany. When I close a tab, it doesn't kill the running plugins
[12:14] <pitti> but there is certainly some fallout from this
[12:14] <mjg59> Firefox is fine
[12:14] <Riddell> pitti: could you give back adept too (for fixed --dist-upgrade)
[12:14] <dendrobates> pitti: I don't want to hold up tribe 4 for it.
[12:15] <seb128> mjg59: bug #125209
[12:15] <ubotu> Launchpad bug 125209 in epiphany-browser "Flash video in Epiphany doesn't die when browser is closed" [Low,Confirmed]  https://launchpad.net/bugs/125209
[12:15] <pitti> Riddell: done
[12:15] <Riddell> thanks
[12:15] <mjg59> seb128: Ah, thanks
[12:15] <pitti> dendrobates: depends on how serious this is deemed
[12:16] <mjg59> seb128: It is indeed the favicon-fallback plugin
[12:16] <mathiaz> pitti: I don't thinks it's such a serious bug.
[12:16] <pitti> ok
[12:16] <dendrobates> pitti: it is an embarrassment, to say the least.
[12:16] <mathiaz> pitti: it doesn't break apache.
[12:17] <mathiaz> dendrobates, pitti: yes. it's just annoying.
[12:17] <pitti> https://launchpad.net/~ubuntu-archive/+subscribedbugs?field.searchtext=sync
[12:17] <pitti> -> I don't see a sync request for apache2 here
[12:17] <mathiaz> dendrobates, pitti: and I can be disturbing for enduser.
[12:18] <pitti> "LaMP"?
[12:19] <mathiaz> pitti: bug 131118
[12:19] <ubotu> Launchpad bug 131118 in apache2 "Sync apache2 from debian unstable (2.2.4-3)" [Undecided,New]  https://launchpad.net/bugs/131118
[12:19] <mathiaz> pitti: I've subscribed ubuntu-main-sponsor.
[12:19] <pitti> mathiaz: ah, good
[12:20] <pitti> mathiaz: "No Ubuntu changes" -> no relevant ones any more
[12:20] <pitti> mathiaz: synced and closed
[12:20] <dendrobates> mathiaz: mdz wants a custom page to load anyway.  We need to get someone to write some php.
[12:20] <cjwatson> pitti: back now. I can have a look at oem-config nowish if I still have time ...
[12:20] <mathiaz> dendrobates: hum... Is php installed by default ?
[12:20] <cjwatson> pitti: I didn't fix rescue mode, and I didn't realise anyone else had ...
[12:20] <pitti> dendrobates: so if you want, I'll speed up the building and build new server ISOs
[12:21] <pitti> cjwatson: wb
[12:21] <mathiaz> dendrobates: it's not that easy anyway...
[12:21] <cjwatson> bdmurray: pong
[12:21] <dendrobates> pitti: I think that would be best.
[12:21] <pitti> cjwatson: indeed I need to rebuild ubuntu ISOs anyway for the oversizedness
[12:21] <mathiaz> dendrobates: there has been a lot of debate over the default apache web page.
[12:21] <pitti> cjwatson: so, fixing those is highly appreciated (either disabling OEM, or even better, actually make it boot)
[12:21] <pitti> cjwatson: does Kubuntu/edubuntu/xubuntu also have that OEM mode?
[12:21] <ajmitch> dendrobates: what sort of thing does he want on the php page?
[12:21] <cjwatson> mikmor2: a few directories are bind-mounted at various points - /cdrom, /proc, /dev, maybe /rofs
[12:22] <Riddell> kubuntu does have oem
[12:22] <cjwatson> kubuntu is certainly supposed to
[12:22] <mathiaz> dendrobates: there is a bug in the debian tracker about it.
[12:22] <pitti> Riddell: do you want live CDs rebuilt for that?
[12:22] <dendrobates> ajmitch: I don't know, for sure.
[12:22] <Riddell> pitti: yeah, may as well
[12:22] <cjwatson> pitti: everything but edubuntu and server has oem-config
[12:22] <mathiaz> dendrobates, ajmitch: it depends what the target is.
[12:22] <pitti> cjwatson: ok, then we at least don't waste the current edubuntu build
[12:23] <ajmitch> mathiaz: I'm assuming this is only for the LAMP setup?
[12:23] <bdmurray> cjwatson: I thought I heard you mention a design decision as to why "vga=" isn't used with grub.  Is that right?
[12:23] <cjwatson> pitti: just rsyncing desktop now
[12:23] <dendrobates> mathiaz: It is on the server tasks page, and mdz asked the status.
[12:23] <mathiaz> ajmitch: well... I don't know exactly what mdz has in mind.
[12:23] <cjwatson> bdmurray: vga= isn't used because that makes vesafb be used which makes suspend/resume a very sad panda
[12:23] <mathiaz> dendrobates: ah ok...
[12:23] <stgraber> pitti, cjwatson: ok, I'll add for Kubuntu as well, what about Xubuntu ?
[12:24] <cjwatson> stgraber: Xubuntu has OEM mode, though I'm not sure anyone cares ...
[12:24] <mathiaz> dendrobates, ajmitch: there has been a lot of debate about what should be put in the default index page.
[12:24] <pitti> stgraber: I don't think that it's worth the trouble for T4 TBH
[12:24] <cjwatson> as in, haven't heard of any OEMs shipping it. it'll just be the same as Ubuntu's anyway
[12:24] <ajmitch> server tasks is assigned to soren, right?
[12:24] <dendrobates> mathiaz: we can discuss it in the next server meeting.
[12:24] <mathiaz> dendrobates, ajmitch: be it for the LAMP install or the default install.
[12:24] <ajmitch> mathiaz: I imagine there would be :)
[12:24] <cjwatson> bdmurray: IIRC because vesafb doesn't know how to reprogram the card's registers on resume
[12:25] <stgraber> pitti: ok, so we'll have it for Ubuntu and Kubuntu now, I can still add others later
[12:25] <mathiaz> dendrobates, ajmitch: every know and then there is a end user that shows up and claims that apache/debian/ubuntu has hacked its server
[12:25] <pitti> mathiaz: lol
[12:25] <ajmitch> mathiaz: they make for funny stories online, but yes, I understand the problem
[12:25] <mathiaz> dendrobates, ajmitch: because their ISP/web hosting has messed up the configuration of the httpd server.
[12:26] <ajmitch> pitti: look up the one about centos
[12:26] <bdmurray> cjwatson: hmm, does that pertain to bug 129910 at all?
[12:26] <ubotu> Launchpad bug 129910 in Ubuntu "tty[1-6]  are active but display nothing in Gutsy" [Undecided,New]  https://launchpad.net/bugs/129910
[12:26] <dendrobates> mathias ajmitch:  I know :) but we can't design to the lowest common denominator.
[12:26] <mathiaz> ajmitch: yeah - that was the last one... last year I think.
[12:26] <cjwatson> bdmurray: yes, it's entirely possible that removing vga= will fix it
[12:27] <dendrobates> pitti: How long will the rebuild take?
[12:27] <cjwatson> bdmurray: it's rather odd that they have locale=es_ES in their boot options
[12:27] <dendrobates> pitti: of the iso?
[12:27] <cjwatson> that's supposed to be filtered out
[12:27] <pitti> dendrobates: apache2 doesn't take that long, but the publishing will
[12:28] <cjwatson> ... oh, bugger
[12:28] <pitti> dendrobates: all in all, two hours maybe
[12:28] <cjwatson> that's definitely a bug, I think ubiquity needs to ship /etc/preseed.aliases
[12:28] <bdmurray> cjwatson: right, they admit as much so I was wondering if it was dismissable
[12:28] <dendrobates> mathiaz: we need to retest either tonight or tomorrow morning.  I need to get my sparc setup anyway.
[12:29] <mathiaz> pitti: ok. So we'll have to start over to cd testing.
[12:29] <cjwatson> bdmurray: we started filtering out vga= from the options set on the installed system in the dapper installer
[12:29] <mathiaz> pitti: when will you release tribe-4 ?
[12:29] <cjwatson> bdmurray: so an upgrade from before then could still have it
[12:29] <cjwatson> bdmurray: it seems like a valid kernel bug though
[12:29] <pitti> mathiaz: tomorrow afternoon (before meeting) if all goes well
[12:29] <mathiaz> dendrobates: I'm not sure if we'll have time tomorrow morning.
[12:29] <mathiaz> dendrobates: morning for us.
[12:30] <cjwatson> bdmurray: vga= is one of the more common options people fiddle with, even if it does break suspend/resume (and of course lots of people don't care about that)
[12:30] <cjwatson> "I want higher resolution on my virtual terminals" is actually a fairly common thing to want if you live at VTs
[12:30] <dendrobates> mathiaz: I can get started pretty early.  We'll do the best we can.
[12:31] <mjg59> cjwatson: Ben made vesafb a builtin. I suspect this may have broken something in the framebuffer startup script - I'm waiting for him to fix it before looking into it too closely
[12:31] <cjwatson> mjg59: builtin? interesting
[12:31] <cjwatson> bdmurray: please reassign to the kernel for now
[12:31] <mathiaz> dendrobates: ok. I'll test the i386 iso later today then.
[12:31] <bdmurray> cjwatson: okay, thanks for all of the information
[12:31] <dendrobates> mathiaz: I have to help feed and put babies to bed.  I'll be back in a few hours to do the amd64 tests.
[12:33] <cjwatson> bdmurray: np, an interesting bug
[12:33] <pitti> ogra, LaserJock, stgraber: http://cdimage.ubuntu.com/edubuntu/daily-live/20070808/ (far from being oversized)
[12:33] <LaserJock> saw it
[12:35] <LaserJock> pitti: what the? I wonder what happened, we dropped es, xh, fr, de lang packs and amd64 dropped 55MB
[12:35] <cjwatson> oh!
[12:35] <cjwatson> I, er, kind of forgot boot=casper
[12:36] <pitti> LaserJock: no idea
[12:36] <LaserJock> pitti: hmm, well at least they're under :-)
[12:37] <stgraber> ogra: want something specifc to be tested on Edubuntu server ?
[12:38] <stgraber> pitti: "panel encountered a problem with loading "OAFIID:GNOME_FastUserSwitchAplet"" known issue ?
[12:38] <pitti> stgraber: no, that worked for me
[12:39] <stgraber> pitti: just got that at first session open time on edu server
[12:41] <cjwatson> pitti: oem> if you F6 it and add boot=casper and splash before --, it should get further
[12:41] <stgraber> ogra, pitti: no fast-user-switching-applet installed on Edubuntu
[12:41] <seb128> stgraber, pitti: that's because the applet has been added to the panel and is not installed
[12:41] <stgraber> ogra, pitti: but enabled on gnome default session, so we have that error
[12:42] <seb128> fast-user-switching-applet should be installed on edubuntu
[12:42] <pitti> darn
[12:42] <pitti> seb128: does it really make sense there? last time you said no
[12:42] <pitti> and on a thin client that sounds reasonable
[12:42] <cjwatson> mikmor2: I just reread; during target-config, I don't think you have any shared directories. Any particular reason for asking?
[12:42] <LaserJock> I don't think it's in the seeds
[12:43] <pitti> LaserJock: no, it's not
[12:43] <seb128> pitti: I didn't think about the panel profile when replying some days ago
[12:43] <seb128> I think it would be easier to add it to edubuntu and to make the applet exit if the config is a ltsp one
[12:44] <cjwatson> evand: do you think it would make sense to disable migration-assistant in ubiquity's oem mode?
[12:44] <cjwatson> I can't imagine an OEM wanting it ...
[12:45] <alex-weej> ok so i installed to my USB disk
[12:45] <pitti> seb128: ah, so you can't tell the panel to load an applet only if it's present?
[12:45] <cjwatson> pitti: desktop oem boot options fixed for your next build
[12:45] <pitti> cjwatson: thanks
[12:45] <cjwatson> pitti: I'm testing it now to see if I can save time
[12:45] <alex-weej> but i foolishly changed the grub installation device from (hd0) to (hd3) thinking that it corresponded to kernel sdd
[12:46] <alex-weej> so ubiquity crashed
[12:46] <alex-weej> or rather gave up with a "fatal error" installing grub
[12:46] <seb128> pitti: no, it'll display an error if it can't load it
[12:46] <cjwatson> it's a known ubiquity bug that it doesn't validate that against grub's device.map
[12:46] <pitti> cjwatson: sorry, can't test right now, I have two installs running already
[12:46] <cjwatson> pitti: not a problem, I'm doing it
[12:46] <alex-weej> cjwatson: how can i get this installed? grub setup refuses to accept (hd0) anyway
[12:46] <pitti> my hd sounds like a sawmill
[12:46] <LaserJock> heh
[12:47] <alex-weej> pitti: record it and make some techno
[12:47] <cjwatson> alex-weej: have a peek in /target/boot/grub/device.map to see what grub thinks the mapping is
[12:47] <alex-weej> cjwatson: /target was unmounted
[12:47] <cjwatson> alex-weej: mount it again
[12:47] <cjwatson> it's all still there
[12:47] <alex-weej> cjwatson: it thinks /dev/sdd is (hd2)
[12:48] <cjwatson> alex-weej: right, there you go
[12:48] <alex-weej> grub> setup (hd2)
[12:48] <alex-weej> Error 12: Invalid device requested
[12:48] <alex-weej> grub> setup (hd2,
[12:48] <alex-weej> Error 21: Selected disk does not exist
[12:48] <cjwatson> alex-weej: did you chroot to /target to do that?
[12:48] <alex-weej> cjwatson: no
[12:48] <cjwatson> do so
[12:48] <cjwatson> there won't be a /boot/grub/device.map in the live session
[12:49] <cjwatson> so grub is unlikely to have device mappings there
[12:49] <cjwatson> at least, they might differ
[12:49] <alex-weej> now grub doesn't know any devices in the chrooted env
[12:49] <cjwatson> mount --bind /dev /target/dev from outside
[12:49] <cjwatson> oh, and mount --bind /proc /target/proc too
[12:49] <cjwatson> (ubiquity does both)
[12:50] <alex-weej> yay
[12:50] <alex-weej> hm. grub> setup (hd2)
[12:50] <alex-weej> Error 12: Invalid device requested
[12:50] <cjwatson> ubiquity hadn't finished with the target system yet though
[12:51] <cjwatson> hadn't removed extra junk packages or copied logs
[12:51] <alex-weej> can i get it to resume?
[12:51] <cjwatson> TBH, it might be easier to start over :-(
[12:51] <cjwatson> no, sorry
[12:51] <alex-weej> aaaaah
[12:51] <alex-weej> lol
[12:51] <alex-weej> ok just to check for sure
[12:51] <cjwatson> we do need to add that validation rather urgently
[12:51] <cjwatson> or at least make it reask rather than bail out
[12:51] <alex-weej> reask for sure
[12:51] <cjwatson> validation's tricky 'cos we don't have a device.map earlier on
[12:52] <cjwatson> (I hate grub. Part #47 in a series of #939)
[12:52] <alex-weej> it's not the first time i've been caught by this. one of my less technical but technical-enough-to-think-he-knows-what-he's-doing friends had a multidisk setup that ubiquity crapped out on
[12:52] <cjwatson> same thing, or something else?
[12:52] <pitti> mathiaz: I invalidated the server CDs now
[12:52] <alex-weej> cjwatson: same thing but not with an external disk
[12:52] <pitti> mathiaz: (in the tracker)
[12:53] <alex-weej> so just to check, i'm installing to sdd, what device string should i give ubiquity to use for grub?
[12:53] <alex-weej> it's defaulting to (hd0)
[12:53] <mathiaz> pitti: ok. I'll wait for the new isos and redo all the test then.
[12:54] <cjwatson> alex-weej: you can give it /dev/sdd, you know ....
[12:54] <stgraber> pitti: About edubuntu, are you going the ISOs with a fix soon or shall I continue testing those ones ? (fix being adding fusa to the seed or find a way to make gnome not load it)
[12:54] <alex-weej> cjwatson: magic.
[12:55] <cjwatson> grub-install notices that and converts
[12:55] <pitti> stgraber: meh, that's going to take two hours at least; not sure wheter we should fix it for Tribe 4
[12:55] <alex-weej> cjwatson: sure there's no sneaky way to get it to skip the file copying? :/
[12:56] <pitti> stgraber: 'make gnome not load it' is out of question for T4, I'm afraid; seeding would be ok, but it needs to go to -desktop (so it takes more like three hours to get new CDs)
[12:56] <stgraber> pitti: the Desktop one being probably affected as well, it'll be the first thing a Tribe-4 users will see on the live environment
[12:56] <alex-weej> when is Tribe 4 due?
[12:56] <pitti> stgraber: oh, right, servers, too?
[12:56] <cjwatson> alex-weej: nope, sorry
[12:56] <stgraber> pitti: I'm on the server one currently
[12:56] <alex-weej> cjwatson: ok, i'll watch some Big Brother Live
[12:56] <pitti> stgraber: please continue testing, though, just to find more serious bugs
[12:57] <cjwatson> alex-weej: tomorrow, or today depending on your timezone
[12:57] <pitti> stgraber: ok, I'll prepare a seed update
[12:57] <alex-weej> cjwatson: ah ffs. only reason i am messing with tribe 3 is to do QA
[12:57] <pitti> seb128: that's just fast-user-switch-applet, right?
[12:57] <cjwatson> alex-weej: http://wiki.ubuntu.com/GutsyReleaseSchedule has the dates
[12:57] <seb128> pitti: yes
[12:57] <alex-weej> suppose i shoulda checked that first.
[12:57] <stgraber> pitti: everything running edubuntu-desktop must be affected (desktop, server desktop and of course thin clients :))
[12:58] <alex-weej> right i'll just get going with that tomorrow then
[12:58] <alex-weej> time to reboot my crufty ex-Edgy upgrade
[01:00] <pitti> stgraber, LaserJock: ok, seeds changed, edubuntu-meta updating, invalidating CD images in tracker
[01:00] <stgraber> ok
[01:00] <LaserJock> pitti: k
[01:01] <stgraber> ogra: we still have a nbd-server crash reported by apport
[01:02] <pitti> stgraber: please file it, but it's too late to fix for Tribe 4
[01:02] <stgraber> ogra: it's the usual init.d script which segfaults (when it shouldn't even be run as we are using inetd ...)
[01:02] <stgraber> pitti: it must already have been filed as we had it with Tribe-3 :)
[01:03] <pitti> stgraber: hm, it wasn't a milestoned bug
[01:03] <stgraber> hmm, so maybe it was fixed and a new version of nbd-server introduced it again, will check
[01:04] <pitti> xubuntu desktop CDs ready for testing
[01:06] <alex-weej> what
[01:06] <cjwatson> pitti: desktop oem seems largely ok except for the wart that autologin and such isn't configured
[01:06] <cjwatson> easy to fix, but not for tribe-4 of course
[01:06] <pitti> cjwatson: great
[01:06] <alex-weej> pitti: is there some kind of ETA on ubuntu tribe 4?
[01:07] <pitti> alex-weej: tomorrow
[01:07] <pitti> cjwatson: if you have a fix right now, it's possible to sneak it in
[01:07] <alex-weej> pitti: it's already 9 Aug in Europe/London :P
[01:07] <pitti> cjwatson: but better not, I figure, if it's not essential
[01:07] <pitti> alex-weej: here, too, but I still have 23 hours :)
[01:08] <cjwatson> pitti: I don't have it right now
[01:09] <alex-weej> pitti: are we just waiting for the images? (i.e. are the archives updated?)
[01:09] <pitti> alex-weej: need to get a fixed apache2 and edubuntu-meta, and rebuild some images
[01:09] <alex-weej> oh ok
[01:10] <cjwatson> hmm, wonder if there's a convenient python equivalent of chroot /target [ -f /etc/gdm/gdm-cdd.conf ] 
[01:11] <cjwatson> s'pose I could just use os.path.lexists
[01:11] <cjwatson> but no, I need the actual path. hmm.
[01:12] <cjwatson> think I'll need to clone-and-hack os.path.realpath
[01:14] <pitti> does anyone feel like writing a tribe 4 wiki page, with the data points from https://lists.ubuntu.com/archives/ubuntu-marketing/2007-August/002167.html?
[01:14] <pitti> apparently Corey doesn't have time, he didn't answer to my mail
[01:16] <stgraber> pitti: OpenOffice 2.3 should also be mentioned imo
[01:16] <pitti> right
[01:16] <pitti> and ubiquity OEM
[01:16] <cjwatson> mm, maybe
[01:17] <cjwatson> still a bit raw, not setting up the oem user correctly
[01:17] <cjwatson> but I guess it's ok to mention, the ubiquity side works fine
[01:17] <cjwatson> you have to know to run oem-config-prepare though
[01:17] <cjwatson> it doesn't tell you and the icon that was supposed to be installed isn't
[01:18] <cjwatson> (already fixed in bzr, I think)
[01:18] <pitti> cjwatson: that should go to the wiki page and release notes then
[01:18] <cjwatson> might be better to announce in tribe-5 instead with fewer caveats?
[01:19] <pitti> cjwatson: WFM (we just have the gfxboot entry now)
[01:19] <cjwatson> yeah
[01:19] <pitti> cjwatson: but we can leave that for the adventurous :)
[01:24] <Riddell> calc: openoffice still doesn't start for me in Kubuntu tribe 4 candidate, despite pitti adding the openoffice.org-gtk larks
[01:25] <Riddell> at least the splash screen is back to blue though :)
[01:25] <pitti> Riddell: yeah, we noticed; that's a release-note thing, I'm afraid
[01:25] <Riddell> pitti: same in ubuntu too?
[01:25] <pitti> Riddell: no, it has always worked there
[01:25] <Riddell> right
[01:25] <pitti> Riddell: installing -gtk and -kde on Ubuntu works
[01:26] <stgraber> why isn't ogra around ...
[01:26] <pitti> Riddell: calc already isolated the responsible library, but AFAIK there's no real fix yet
[01:26] <pitti> stgraber: he wanted to go to bed early and get up early
[01:26] <Amaranth> alex-weej: Can you be less... hostile in your bug reports?
[01:26] <alex-weej> Amaranth: excuse me?
[01:26] <stgraber> "This workstation isn't authorized to connect to server" ...
[01:28] <Amaranth> alex-weej: 'Nonsense!'
[01:28] <pitti> stgraber: ?
[01:28] <alex-weej> Amaranth: i can if you tell me what i've said that seems "hostile"
[01:28] <alex-weej> Amaranth: ah come on, that's not hostility :/
[01:28] <stgraber> pitti: error message I get when trying to login from a thin client
[01:28] <Amaranth> alex-weej: That bug is marked as triaged, that means we know what the problem is.
[01:28] <pitti> stgraber: that sounds...serious
[01:28] <Amaranth> alex-weej: No need for a 'me too' post anyway.
[01:28] <stgraber> pitti: I'm rebuilding the thin client image adding a root passwd so I can see what actually happens
[01:28] <alex-weej> Amaranth: i was adding information. it's a two pixel no-mans land.
[01:29] <stgraber> pitti: it's better than with previous ldm, at least we have an error message :) (even if it doesn't really tell us anything interesting)
[01:29] <Riddell> calc: I installed openoffice.org-style-tango openoffice.org-gnome openoffice.org-base and now it start up
[01:29] <Riddell> not sure which one of those is actually responsible
[01:29] <pitti> Riddell: -gnome probably
[01:30] <pitti> Riddell: can you uninstall -tango to check? -base should be there anyway, I figure?
[01:30] <Amaranth> alex-weej: I'll fix it tomorrow
[01:30] <alex-weej> Amaranth: i guess you know how much i hate compiz anyway so that probably tainted your perception of the tone of my comment a little bit :P
[01:30] <pitti> Riddell: or don't you install it deliberately?
[01:30] <Amaranth> alex-weej: Could be.
[01:31] <Riddell> pitti: does indeed seem to be openoffice.org-gnome
[01:32] <Riddell> pitti: -base we stopping including a couple releases ago in favour of kexi which is much better, but then we removed kexi lsat release since it was taking up space
[01:33] <LaserJock> pitti: the Edubuntu Server Addon CD didn't need to be rebuilt I don't think, are the Server and Server Addon CDs built separately
[01:33] <pitti> LaserJock: no, it's one step
[01:33] <LaserJock> k, just wondered
[01:33] <stgraber> pitti: hmm, this time it let login ...
[01:34] <pitti> stgraber: heisenbug?
[01:34] <LaserJock> pitti: how long before Edubuntu Desktop gets rebuild, roughly?
[01:35] <pitti> LaserJock: about 90 minutes until live is ready, 60 for server
[01:35] <pitti> LaserJock: oh, add 15 minutes, current publisher isn't finished yet
[01:35] <LaserJock> pitti: hmm, I wondered it it'd be worth it to add back in at least one or two lang packs
[01:36] <stgraber> pitti: but ldm didn't reload after logout (so black screen after logout) ...
[01:36] <LaserJock> but maybe it's just as well to leave that until after Tribe 4
[01:36] <LaserJock> I don't want to waste a build run on miscalculating lang packs
[01:37] <pitti> LaserJock: can do
[01:37] <pitti> LaserJock: I have my script here, and I'll be cautious; which langs are the most important for you?
[01:37] <LaserJock> well, we had es de fr xh
[01:38] <LaserJock> on amd64
[01:39] <LaserJock> I want to say es and fr, but maybe if I leave out de ogra'll be mad at me ;-)
[01:39] <Kmos> and pt ?
[01:40] <Kmos> it's one of the most talked in the world
[01:41] <stgraber> pitti: ok, found the actual bug with ldm, it doesn't disconnect from SSH on session logout, so it doesn't reload and I'm stuck on a nice black screen :)
[01:42] <LaserJock> pitti: if you can do es de fr on amd64 I think that should be good, I'm not sure how much the fast-user-switch stuff will take but I figure it should work out
[01:42] <pitti> LaserJock: I leave the server CDs alone; i386 is full, and amd64 not that important for tribe, I figure
[01:42] <stgraber> pitti: killing ssh by hand seems to be the fix
[01:43] <pitti> LaserJock: dependencies look like you already have everything
[01:43] <pitti> LaserJock: and it adds 600 kB
[01:44] <LaserJock> pitti: right
[01:45] <stgraber> LaserJock: Were you at last edubuntu meeting (today) ? Has ogra said something about ldm2 (like a new version coming but not uploaded yet) ?
[01:45] <pitti> LaserJock: adding es will fill up i386
[01:46] <pitti> LaserJock: you can have one more
[01:46] <LaserJock> stgraber: no, it's 5am for me
[01:46] <pitti> LaserJock: (on amd64)
[01:46] <LaserJock> pitti: de I guess :-)
[01:46] <pitti> LaserJock: depends on your target audience, I don't know
[01:46] <pitti> LaserJock: .de isn't a particularly popular edubuntu target, I think
[01:47] <LaserJock> pitti: ok, fr then, although I think many of our testers know German
[01:47] <pitti> LaserJock: pt?
[01:47] <LaserJock> well, we had es, fr, de, and xh before
[01:48] <LaserJock> I don't know if I should go beyond that
[01:48] <Kmos> pt is more talked than es or fr
[01:48] <pitti> no, just one more
[01:48] <Kmos> brazil talks pt, spanish people understand pt, portugal
[01:48] <LaserJock> pitti: I mean, beyond the langs that were already there, I was just picking amongst the old ones, so no pt
[01:48] <Kmos> and so many other countries
[01:48] <Kmos> africa
[01:49] <LaserJock> I honestly don't think it's going to matter for Tribe4
[01:50] <pitti> LaserJock: ok, just adding spanish now
[01:54] <pitti> second publisher running now, for apache+edubuntu-meta binaries
[01:55] <mathiaz> pitti: kwel. Server isos should be ready in an hour or so ?
[01:55] <pitti> Riddell: do you actually want new Kubuntu CDs for OEM when we don't advertise it for Tribe4 yet?
[01:55] <pitti> mathiaz: yes
[01:55] <cjwatson> Kmos: our standard ordering was calculated from published speaker numbers as far as I know, and es, bn, hi, ar are all ahead of pt there
[01:55] <cjwatson> Kmos: Edubuntu typically deviates from our usual ordering though, since it has a weird distribution
[01:56] <ajmitch> hence xh?
[01:56] <cjwatson> Kmos: we do not want to get into language wars
[01:56] <pitti> cjwatson: new standard ordering is en, es, xh, pt, de, fr, bn, hi, ar, ru, zh, though, due to considering input support
[01:56] <Riddell> pitti: what's broken about it currently?
[01:56] <pitti> Riddell: just the OEM gfxboot thing
[01:56] <cjwatson> ajmitch: xh is an anomaly because Canonical paid for the translation ages ago, but since it's so out of date the language pack size is tiny and it hardly matters
[01:57] <cjwatson> pitti: oh. update http://people.ubuntu.com/~pitti/scripts/langpacksize please :-)
[01:57] <cjwatson> (preferably with a justification so we remember)
[01:57] <pitti> cjwatson: I guess I should replace it with a symlink to current bzr head on codebrowse or so :)
[01:58] <LaserJock> ah, that would be nice, I was using it this morning :-)
[01:58] <Riddell> pitti: lets not bother then
[02:00] <pitti> LaserJock, cjwatson: updated my script on people for now
[02:00] <Kmos> cjwatson: ok..
[02:01] <Kmos> I think the most spoken, are the first.. but that not the rules here.
[02:04] <cjwatson> Kmos: much harder to judge objectively though
[02:05] <cjwatson> bedtime
[02:05] <Kmos> cjwatson: good night
[02:09] <pitti> bye cjwatson
[02:10] <LaserJock> pitti: you gonna be up until the new disks are done?
[02:11] <pitti> LaserJock: probably, unless I fall off my chair :)
[02:11] <pitti> gosh, I'm sitting in front of that thing for 20 hours now
[02:11] <sistpoty> gn8 everyone
[02:11] <pitti> bye sistpoty
[02:11] <bryce> pitti :-/
[02:12] <sistpoty> gn8 pitti, and good luck with staying awake ;)
[02:23] <pitti> iwj: heh, the fsck bug you reported is the one I asked you to take a look at recently :)
[02:29] <pitti> yay, I removed 2 MB langpacks from ubuntu desktop amd64, and got 4 MB more free
[02:30] <pitti> ubuntu desktop CDs available for testing
[02:33] <ArneGoetje> pitti: finally, eh? now you can get some sleep?
[02:33] <pitti> ArneGoetje: not yet :/
[02:34] <pitti> we still need new server CDs, and the entire range of edubuntu
[02:34] <ArneGoetje> pitti: and all today?
[02:34] <pitti> and due to an inadvertent key press from me triggering another ubuntu livefs build, I need to wait for half an hour before I can start the edubuntu ones
[02:34] <pitti> ArneGoetje: yes
[02:35] <bryce> pitti, congrats :-)
[02:35] <Riddell> stgraber: if I'm looking at the results of a Kubuntu CD in isotesting, then click on "build list" it takes me back to the Ubuntu list rather than the Kubuntu list
[02:36] <bryce> (er, for getting the desktop CD's ready, shucks about the extra keypress, guess that happens after 20+ hrs)
[02:37] <pitti> bryce: the new livefs build won't manifest itself in new ISOs on cdimage, don't worry :)
[02:37] <pitti> it's just wasted time
[02:38] <mathiaz> pitti: which is very annoying after 20hrs of sitting in front of your computer...
[02:38] <Riddell> pitti: all 4 kubuntu CDs work, except for known bugs (live verify CD broken, openoffice broken)
[02:39] <pitti> Riddell: ok, good
[02:39] <pitti> ... ish
[02:39] <pitti> Riddell: verify is curious
[02:39] <StevenK> Openoffice is still broken?
[02:39] <pitti> workaround with adding -gtk wasn't enough
[02:39] <pitti> it needs -gnome actually
[02:39] <StevenK> Neat. :-(
[02:39] <pitti> but that carries heavy dependencies, I figure
[02:40] <pitti> Riddell: please unseed ooo-gtk from kubuntu, before we forget about it
[02:40] <Riddell> ok
[02:43] <pitti> meh, on the ubuntu live, deskbar applet immediately crashes and gives bug-buddy
[02:43] <pitti> I didn't see that before
[02:52] <pitti> mathiaz: new ubuntu server isos up
[02:53] <mathiaz> pitti: great. thanks.
[02:54] <pitti> can anyone please test the latest desktop and check whether deskbar-applet crashes right at start?
[02:54] <pitti> dendrobates: wb
[02:54] <pitti> dendrobates: new server images just arrived
[02:54] <dendrobates> pitti: great timing
[02:55] <dendrobates> pitti: I'll get to testing
[02:55] <dendrobates> er.. downloading. :)
[02:55] <pitti> stgraber: new edubuntu servers up
[02:55] <pitti> ogra: ^
[02:57] <mikmor2> cjwatson: Are you awake by chance?
[02:57] <pitti> mikmor2: not any more
[02:58] <mikmor2> do'h.
[02:58] <ArneGoetje> pitti: where can I download the new images?
[02:59] <pitti> ArneGoetje: the iso tester tracker has links
[03:00] <pitti> ArneGoetje: if you already have older CDs, you probably want to use rsync or jigdo, unless you have insane bandwidth :)
[03:00] <ArneGoetje> pitti: where can I find it?
[03:00] <mathiaz> ArneGoetje: https://iso.qa.stgraber.org/isotesting/
[03:00] <pitti> (topic)
[03:00] <ArneGoetje> mathiaz: thanks
[03:01] <pitti> ArneGoetje: thanks for testing
[03:01] <pitti> ArneGoetje: iso tracker has rsync commands etc, too
[03:01] <mathiaz> ArneGoetje: this is the website used to track the testing effort.
[03:03] <pitti> ok, edubuntu lives are building
[03:05] <evand> cjwatson: I'm fine with that
[03:06] <dobey> if i want to rebuild the kernel on my system, with a couple of extra patches, what's the preferred method for that? patching the source in /usr/src/linux* and doing make kpackage or whatever it is?
[03:12] <iwj> pitti: Indeed, I remember the question.  But I couldn't remember if it was actually a bug report so I thought I'd file it in case it wasn't.
[03:12] <iwj> Hope that's not too annoying.  I'll take a look at it properly after FF.
[03:13] <pitti> iwj: not at all
[03:13] <ArneGoetje> pitti: can I use the rsync script for the daily builds? Or is that different from iso tracker?
[03:13] <pitti> ArneGoetje: yes, rsync works well for both desktops and also alternates
[03:13] <pitti> ArneGoetje: the iso tracker gives you the correct rune; if the local file name does not match, you have to specify it instead of '.', of course
[03:14] <ArneGoetje> pitti: ok
[03:15] <pitti> so, this deskbar applet thing still crashes very often for me
[03:15] <pitti> I'll try the live CD on real iron and then fall to bed
[03:15] <pitti> cu tomorrow, and happy testing so far
[03:15] <pitti> wb LaserJock
[03:15] <pitti> LaserJock: edubuntu server is there, desktop ETA 20 minutes
[03:15] <pitti> bye everyone
[03:15] <LaserJock> alright
[03:15] <mathiaz> pitti: see ya.
[03:15] <iwj> pitti: Sleep well.
[03:15] <LaserJock> thanks pitti
[06:55] <LaserJock> anybody know what "Printing Notification Icon" is?
[06:56] <RAOF> Isn't that the print queue icon for the notification area?
[06:58] <fabbione> morning everybody
[06:59] <ion_> ning
[06:59] <LaserJock> RAOF: but why is it in the menu?
[06:59] <LaserJock> that's kinda weird
[07:02] <calc> Riddell: it starting or not seems to be due to when gtk gets properly initialized which is probably why it sometimes works for kubuntu users and sometimes not
[07:02] <LaserJock> morning fabbione
[07:02] <calc> Riddell: it looks like libvclplug_gtk680li.so probably does proper gtk initialization and allows ooo to run on some systems
[07:03] <RAOF> LaserJock: Absolutely no idea.
[07:39] <calc> http://kerneltrap.org/node/14148 <- any thoughts about having ubuntu default to noatime,nodiratime ?
[07:39] <calc> claims in the thread are that it can speed up disk io by a huge amount on things like kernel compiles, etc
[07:40] <ScottK> calc: "Makes kernel compiling faster" doesn't sound like a common enough use case to support by default....
[07:42] <calc> ScottK: the claim specifically for kernel compile time was 40% increase in speed
[07:42] <calc> they don't include much in the way of benchmarks in the thread other than that part
[07:43] <RAOF> Actually, they do.
[07:43] <calc> "People just need to know about the performance differences - very few
[07:43] <calc> realise its more than a fraction of a percent. I'm sure Gentoo will use
[07:43] <calc> relatime the moment anyone knows its > 5% 8)"
[07:43] <calc> relatime actually writes out atime still but in a different manner, turning it off entirely does even less io
[07:43] <calc> RAOF: ah maybe i haven't read that far yet
[07:43] <RAOF> A guy with a thinkpad sees ~10-20% speed increase on kernel compile, from memory.
[07:44] <calc> its a fairly long page
[07:44] <RAOF> Indeed.
[07:44] <calc> hmm ingo claims ubuntu already does that
[07:44] <calc> but its not in my fstab
[07:45] <ScottK> You got the special "Make calc wait on OOO" edition.
[07:57] <ScottK> calc: I guess I'd want to understand better why someone would care about atimes before thinking to seriously about changing the default.
[08:02] <keescook> well that's ugly.  nanosleep doesn't return while in the initramfs
[08:06] <doko_> good morning
[08:07] <LaserJock> morning doko
[08:27] <ArneGoetje> doko: moin
[08:27] <doko> ArneGoetje: moin
[08:31] <ArneGoetje> doko: about the sun-java6-bin issue...
[08:32] <ArneGoetje> I think it would be best if we path the fontconfig config file provided in the package and replace the Japanese font entry of IPAMona with SazanamiMincho. just like in the sun-java5-bin package.
[08:34] <doko> ArneGoetje: ok, will do that, probably not this week. more interesting is that we make these changes to the coming icedtea/openjdk
[08:34] <ArneGoetje> doko: ?
[08:34] <doko> ArneGoetje: the free "sun-java7"
[08:35] <ArneGoetje> doko: ah...
[08:35] <ArneGoetje> doko: personally I currently use the java5 package, because java6 doesn't work with my java applications
[08:38] <ArneGoetje> doko: but one more change we could probably make to the fontconfig file in the sun java packages: for the Korean fonts, it currently referrs to the Baekmuk Batang font in the ttf-baekmuk package. However, the UnBatang and UnBatangBold fonts from the ttf-unfonts packages are 1. smaller in size and 2. more beautiful. :)
[08:40] <doko> ArneGoetje: are those free as well?
[08:40] <ArneGoetje> GPL
[08:48] <Hobbsee> greetings all
[09:06] <Hobbsee> geser: no.  i just hadnt gone to bed by that point.
[09:11] <Mithrandir> Hobbsee: "Girls don't exist on the internet"> excellent article, but quite sad.
[09:12] <Hobbsee> Mithrandir: :D
[09:12] <Hobbsee> Mithrandir: indeed
[09:12] <Hobbsee> Mithrandir: i find it's useful to give out to nutters from userland irc channels.
[09:13] <Hobbsee> my point there was - "is there any point in notifying people abotu it all the time, or ignoring the lesser offenses"
[09:13] <rosegrass> A package was in Ubuntu has version "20060101-1" but Debian now maintain it too and use "0.0.20070101-1" for version number... How Ubuntu sync this package?
[09:13] <ion_> hobbsee: Man, that must be irritating, sir.
[09:13] <Hobbsee> ion_: :P
[09:14] <StevenK> rosegrass: We can't, we'd need to add an epoch.
[09:14] <Mithrandir> Hobbsee: dude, that must be strange, yes.
[09:14] <ion_> If the Ubuntu version is 20060101-1, it has been synced from Debian, or Debian has taken it from Ubuntu in the first place, right?
[09:14] <Hobbsee> Mithrandir: at least no one manages to make that mistake in person, though
[09:15] <rosegrass> ion_, no, it's not true. 20060101-1 is only in Ubuntu.. And some new person repacks it in Debian..
[09:16] <rosegrass> StevenK, does that mean we'll always manually add an epoch??
[09:16] <rosegrass> StevenK, if there's new Debian version...
[09:17] <StevenK> The problem is 20060101-1 is greater than 0.0.20070101-1
[09:17] <rosegrass> StevenK, yes..
[09:17] <StevenK> So, yes, we'd need to add an epoch, or convince the Debian maintainer to do it.
[09:19] <stdin> thing is, debian packages guidelines say that a date as a version should be something like 0.0.date in case upstream make a 1.0 relase
[09:19] <rosegrass> stdin, yes.. it is..
[09:21] <ion_> Perhaps even 0~date in case theres going to be a 0.0.1 :-)
[09:22] <Mithrandir> ion_: 0~ means we can't sync as well
[09:22] <ion_> mithrandir: Generally?
[09:23] <Mithrandir> ion_: yes, since 0~ is < 0, which is what the sync script compares against.
[09:25] <ion_> mithrandir: Okay
[09:26] <Hobbsee> hmmm.  new ooo without -gnome and -gtk still doesnt work
[09:27] <Hobbsee> ah, works if you add -gnome
[09:28] <StevenK> Hobbsee: But -gnome probably drags in a shedload of dependancies.
[09:28] <Hobbsee> StevenK: some, yes.  it's been force added to the kubuntu cds, as there's no other fix coming quickly
[09:29] <Hobbsee> calc: ^
[09:29] <StevenK> Hobbsee: Twitch.
[09:30] <Hobbsee> drat.  it hasnt
[09:30] <Hobbsee> only -gtk has, and that doesnt fix the issue.
[09:31] <infinity> I thought it had been confirmed that -gtk fixed it...
[09:31] <StevenK> I thought Riddell said it hadn't fixed it.
[09:31] <Hobbsee> infinity: it seems not.  on the bug report, a copule of people mentioend it, but not everyone found that to be the case
[09:32] <infinity> Hobbsee: Bleh.
[09:36] <Hobbsee> infinity: i take it you dont care enough to respin the livefs', and the kubuntu cds?
[09:38] <Hobbsee> (or even think about it?)
[09:40] <infinity> Hobbsee: I could do.
[09:40] <Hobbsee> morning pitti!
[09:40] <infinity> Hobbsee: But we'd need a seed change and meta update first, no?
[09:40] <Hobbsee> infinity: ah, now you can leave that decision to pitti
[09:40] <infinity> Well, a seed change and publisher run.  The livefs builds run based on tasks, not metapackages.
[09:40] <Hobbsee> infinity: true that - i'm still reading backscroll and such - and we wouldnt need that meta update and seed change if we werent going to respin
[09:40] <pitti> Good morn*yawn*ing
[09:41] <infinity> I hear that yawn.
[09:41] <infinity> I think I'm going to have a bit of a siesta for an hour or two before I get back to work.
[09:41] <pitti> Hobbsee: new CD images? *cry*
[09:41] <Hobbsee> pitti: right then :)
[09:42] <infinity> pitti: kubuntu/openoffice still busticated, apparently.
[09:42] <infinity> pitti: -gtk wasn't enough, also need -gnome
[09:42] <pitti> infinity: right, we noticed yesterday
[09:42] <pitti> yeah, it worked on ubuntu with just -gtk and -kde
[09:43] <pitti> but -gnome pulls in a lot of more dependencies
[09:43] <Hobbsee> pitti: but gconf, etc, would normally be installed on ubuntu anyway
[09:43] <pitti> probably
[09:43] <pitti> calc identified the library that needs to be fixed
[09:43] <Hobbsee> then, on the other hand, how many people actually use ooo on kubuntu?
[09:43] <pitti> but we don't have a patch yet, and no time to rebuild OO.o
[09:43] <pitti> Hobbsee: Riddell did a lot of CD tests yesterday, and apart from that they were ok
[09:44] <pitti> Hobbsee: so we'll release-note the OO.o issue and give it up for this tribe
[09:45] <pitti> hey seb128
[09:45] <seb128> hello pitti
[09:45] <pitti> seb128: got my sms?
[09:46] <pitti> seb128: in vmware I saw deskbar-applet crashing a lot on a tracker issue (bug-buddy+nonfunctinoal applet), but it doesn't happen on real iron for me
[09:46] <pitti> yay race conditions
[09:46] <pitti> seb128: it's a release-notes thing in any case, but I was curious if you saw that somewhere?
[09:46] <seb128> pitti: ups, didn't notice the sms before now
[09:46] <seb128> no, deskbar works fine for me :/
[09:47] <seb128> and I don't think we got a bug about it
[09:47] <seb128> do you have a backtrace?
[09:47] <pitti> seb128: yeah, bug-buddy produces one
[09:47] <pitti> seb128: I was too tired yesterday, I'll file a bug about it today
[09:47] <seb128> waouh, you worked late
[09:49] <pitti> We urgently need to find someone to flesh out https://wiki.ubuntu.com/GutsyGibbon/Tribe4
[09:51] <Hobbsee> pitti: yeah, cd tests look OK.  ubuntu i386 looks completely untested.  then again, we dont have time to fix anything now anyway
[09:51] <pitti> yeah, but we need more desktop tests
[09:51] <pitti> I'll do the amd64 ones now
[09:51] <pitti> but I need someone for i386, too
[09:52] <Hobbsee> (desktop)
[09:52] <Hobbsee> are they the ones in current/ ?
[09:52] <pitti> Hobbsee: yes
[09:53] <Hobbsee> cool.  grabbing them now
[09:53] <pitti> meh, the guys were so eager to get edubuntu CDs yesterday, and now server doesn't have a single test
[09:54] <ogra> pitti, running
[09:55] <pitti> ogra: ah, thanks!
[09:59] <ogra> pitti, apparently stgraber did a test last night, you guys talked about it in my backlog ;) seems he didnt add the results to the tracker
[10:01] <pitti> seb128: bug 131247
[10:01] <ubotu> Launchpad bug 131247 in tracker "libdeskbar-tracker crashes at session start" [High,New]  https://launchpad.net/bugs/131247
[10:01] <pitti> ogra: yeah, and he tested the previous version, too; he had some trouble with 'not permitted to login' and debugged it, did he mail/bug you?
[10:01] <Amaranth> pitti: heh, that's a fun one
[10:02] <ogra> i got a pm
[10:03] <seb128> pitti: looks like the one where upstream is getting load of duplicates and which should have been fixed in 0.6
[10:03] <pitti> seb128: can you please flesh out the Gnome/Tracker/fusa section on the wiki?
[10:03] <pitti> seb128: oh, known already then? good
[10:03] <seb128> pitti: let me grab some coffee and I'll have a look to it
[10:03] <pitti> seb128: 'trackerd' exiting doesn't point to the actual bug in tracker, though, I figure
[10:03] <ogra> pitti, for the nbd-server issue he had i changed bug 122544 to a sync request
[10:03] <ubotu> Launchpad bug 122544 in nbd "please sync nbd-server from unstable to fix SIGSEGV" [Medium,In progress]  https://launchpad.net/bugs/122544
[10:04] <pitti> ogra: alright
[10:04] <seb128> doesn't look like, no
[10:04] <pitti> ogra: how bad is this? need new CDs for Tribe4?
[10:04] <pitti> seb128: merci
[10:04] <seb128> de rien ;)
[10:04] <seb128> brb, coffee
[10:04] <ogra> pitti, nah, it throws up an apport win on first login after booting, thats all
[10:05] <pitti> ogra: as you wish; I sync it now anyway; can we override the ubuntu changes?
[10:05] <ogra> pitti, we dont use nbd-server in standalone mode, only from inetd ... it crashes only in standalone mode
[10:06] <ogra> yeah, can all be dropped
[10:06] <ogra> wouter is a great upstream
[10:07] <pitti> synced
[10:07] <ogra> thanks :)
[10:08] <asac> or is there known brokeness in current?
[10:08] <ogra> i havent seen the login error he saw at all yet and apparently he hasnt seen it a second time either
[10:09] <pitti> asac: some warts, but not fatal ones
[10:09] <ogra> pitti, ... and the logout thing is serious but not worth a new build imho (this is still called a development relese :) )
[10:09] <Hobbsee> pitti:
[10:09] <Hobbsee> pitti: got a preference for real HW, or a VM?
[10:09] <ogra> *release
[10:09] <pitti> Hobbsee: real hw always :)
[10:10] <Hobbsee> pitti: how am i supposed to pretend that i'm doing my assignmetn, with 2 computers running?
[10:10] <pitti> Hobbsee: one test on real iron, rest on vmware is good
[10:10] <pitti> :)
[10:10] <tepsipakki> how can I find an old bug # from launchpad?
[10:10] <pitti> tepsipakki: advanced search, tick bug states
[10:10] <pitti> tepsipakki: or try google with 'site:launchpad.net'
[10:10] <Hobbsee> hi jono
[10:11] <tepsipakki> um, I mean how to map a malone bug to lp?
[10:11] <tepsipakki> since the numbers don't match
[10:12] <jono> hi Hobbsee
[10:13] <pitti> cjwatson: I filed bug 131250 just for the record; other than that, ubiquity OEM worked well
[10:13] <ubotu> Launchpad bug 131250 in ubiquity "OEM install should have the 'prepare for end user' icon on desktop" [Medium,New]  https://launchpad.net/bugs/131250
[10:14] <tepsipakki> uh, wrong bug number on the changelog :P
[10:15] <Hobbsee> pitti: so far, no luck on poking people into writing release notes
[10:15] <Hobbsee> pitti: i thougth we had a marketing guy - perhaps, wake him up?  :)
[10:16] <cjwatson> calc: Ingo made a mistake, we don't
[10:16] <cjwatson> pitti: huh, now I have to go adjust the patch I already committed ;)
[10:17] <pitti> cjwatson: hm? for OEM?
[10:17] <pitti> cjwatson: (sorry, my brain is not yet fully functional, I'm afraid)
[10:18] <cjwatson> pitti: yes
[10:18] <pitti> Hobbsee: I got an email from Corey that he wanted to start working on it, but he's to bed now, I guess
[10:18] <Hobbsee> pitti: i'd expect so, yes.
[10:18] <pitti> cjwatson: how did you design it?
[10:18] <cjwatson> pitti: don't understand?
[10:19] <cjwatson> it was already meant to be there, it was just a straight bug that it wasn't
[10:19] <pitti> cjwatson: 'I have to go adjust the patch...'
[10:19] <cjwatson> the debian/changelog bit of the patch, to close the bug
[10:19] <pitti> cjwatson: oh, I see
[10:19] <pitti> I just want the bug as a reference, in case we want it in the release notes
[10:20] <cjwatson> ok
[10:21] <Hobbsee> pitti: got called to dinner, but i'll test gutsy i386 desktop when i come back.
[10:22] <pitti> Hobbsee: thanks
[10:26] <doko> pitti: please accept ecj (re-add a missing dependency for ecj-gcj on ecj, causing build failures)
[10:39] <pitti> doko: done
[10:52] <loonyxp> Hello, I have to patch the xorg-server in order to have the default resolution for my flatpanel monitor. Unfortunately it is only possible with the server release version of 1.3.0, since there I can apply mode patches. I have tried to build xorg-server within feisty fawn, but I got an error while building (ld returning status 1) saying "make error [Xdmx] ". I have the dependencies installed that are listed in lauchpad. Has anbody a hint what to do
[10:54] <stdin> make sure to get all the build-dependencies with, apt-get build-dep package
[10:55] <loonyxp> stdin: ah, I'll try that. I am new to ubuntu (and debian), so I didn't know that. Thanks..
[11:02] <pitti> I'm offline for a bit to test-reinstall my desktop, bbl
[11:08] <Kopfgeldjaeger> whats the best way to convert flv videos to mpeg (with ffmpeg or mencoder)?  ffmpeg -f mp4 works, but there is a quality loss...
[11:09] <stgraber> morning
[11:09] <ogra> hey stgraber
[11:09] <ogra> thanks for the bug pointers :)
[11:09] <ogra> (nbd-server is synced)
[11:10] <stgraber> np, did you find a workaround for the ldm bug ?
[11:10] <ogra> no
[11:10] <Chipzz> Kopfgeldjaeger: this really isn't the channel for such questions; no support here
[11:10] <ogra> i'll leave that to sbalneav ...
[11:10] <loonyxp> stdin: so, I tried it, but apt says, there is nothing to update. (I have deb and deb-src entries for gutsy in my sources list)
[11:10] <Kopfgeldjaeger> Chipzz: hm, ok
[11:11] <stdin> loonyxp: hmm, maybe try asking on #xorg they should know be able to help you
[11:12] <Chipzz> Kopfgeldjaeger: your question yesterday about C++/glade coding would be best answered in #gnome or #gtk+ on irc.gnome.org (/msg burito invite <#channel> to get an invite)
[11:12] <Kopfgeldjaeger> Chipzz: cool, thanks
[11:13] <Chipzz> (apart from the fact that c++ has a seperate load of issues, but that's even more off-topic)
[11:13] <loonyxp> stdin: hm, I think it is not an xorg problem, it is a problem of the ubuntu patches, since I was able to build and run the server on another system (gentoo) without the ubuntu patches... who is the maintainer of the xorg-server package for gutsy?
[11:14] <stdin> loonyxp: "Maintainer: Ubuntu Core Developers <ubuntu-devel-discuss@lists.ubuntu.com>"
[11:15] <Hobbsee> with any luck, the other machine wont catch fire.
[11:15] <loonyxp> stdin: thanks
[11:17] <Hobbsee> seb128: do you know about a deskbar application crash?
[11:17] <stgraber> current edubuntu-server is 20070808.1 + fusa added to edubuntu-desktop right ?
[11:17] <seb128> Hobbsee: which one?
[11:17] <stgraber> (so I can post my test results - the fusa can't be loaded bug)
[11:18] <ogra> i think so
[11:18] <Hobbsee> seb128: in moduleloader.py
[11:18] <seb128> Hobbsee: I know about some crashes, your description lacks detail to reply though
[11:18] <Hobbsee> seb128: this is true.
[11:19] <seb128> Hobbsee: https://bugs.launchpad.net/ubuntu/+source/tracker/+bug/131247 ?
[11:19] <ubotu> Launchpad bug 131247 in tracker "libdeskbar-tracker crashes at session start" [High,New] 
[11:19] <Hobbsee> oh, i have no network connectoin anyway
[11:25] <Hobbsee> seb128: oh, bug buddy sends crashes to gnome now.  gnome bug 464978
[11:25] <ubotu> Gnome bug 464978 in general "crash in Deskbar: running the live cd" [Critical,Unconfirmed]  http://bugzilla.gnome.org/show_bug.cgi?id=464978
[11:25] <seb128> Hobbsee: right, and there is no useful information nor backtrace, I'm going to close it
[11:25] <Hobbsee> seb128: there was one here.  i dont know why it didnt send
[11:25] <Hobbsee> seb128: but please do
[11:27] <stgraber> ogra: bug 131259
[11:27] <ubotu> Launchpad bug 131259 in ltsp "[Gutsy]  LDM doesn't disconnect ssh on logout" [Undecided,New]  https://launchpad.net/bugs/131259
[11:28] <ogra> stgraber, thanks
[11:31] <Hobbsee> seb128: apologies.  http://wedontsleep.org/~sarah/deskbarcrash
[11:31] <Hobbsee> seb128: that's happening on 2 machines here
[11:31] <seb128> Hobbsee: looks like the bug I pointed to you ;)
[11:31] <seb128> k, so it's known
[11:31] <Hobbsee> seb128: great.
[11:32] <Hobbsee> seb128: erm, that's the one i just reported, i think
[11:32] <pitti> Hobbsee: I reported it, too
[11:32] <seb128> Hobbsee: <seb128> Hobbsee: https://bugs.launchpad.net/ubuntu/+source/tracker/+bug/131247 ?
[11:32] <ubotu> Launchpad bug 131247 in tracker "libdeskbar-tracker crashes at session start" [High,New] 
[11:32] <Hobbsee> seb128: oh, that one.  right
[11:37] <Hobbsee> pitti: i dont suppose it matters, but NM seems to be segfaulting on shutdown, and is spewing that across the console
[11:37] <pitti> Hobbsee: not for releasability, but bug should still be reported
[11:38] <ogra> Hobbsee, lets get rid of the console then ;)
[11:38] <Hobbsee> ogra: :P
[11:38] <Hobbsee> ogra: i thought nautilus mounted things as sync now.
[11:38] <Hobbsee> not async
[11:38] <ogra> nautilus doesnt mount things :)
[11:39] <ogra> thats gnome-mount
[11:39] <ogra> which is using what pitti gives it i guess :)
[11:39] <Hobbsee> whatever.  gnome in genearl
[11:39] <pitti> Hobbsee: sync? no, why should it, that's evil
[11:40] <Hobbsee> pitti: so i can pull the usb stick out, and still have my data
[11:40] <ogra> ugh ...
[11:40] <coNP> USB sync is mounted async even in Windows...
[11:40] <coNP> s /sync/stick/
[11:41] <ogra> seb128, i think we should make fusa respect the LTSP_CLIENT variable ... i cant imagine its a good idea on thin clients :)
[11:41] <seb128> ogra: what I was saying yesterday, it's a bit later for the tribe now though
[11:41] <ogra> yeah
[11:41] <ogra> well, i need to test it
[11:42] <ogra> but we have setups with 2000 users on ldap out there
[11:42] <seb128> ogra: do you desactivate the "switch" action as well?
[11:42] <ogra> not yet
[11:42] <ogra> i'll do some testing next week what works and what doesnt first ...
[11:42] <ogra> if functionallity works i'm fine to keep it
[11:42] <seb128> ok
[11:42] <seb128> let me know
[11:43] <ogra> but i assume fusa will show all available users ...
[11:43] <ogra> so in ldap setups thats not a god idea
[11:43] <ogra> (if tehy are huge enough)
[11:44] <pitti> Hobbsee: 'sync' isn't the right answer, but http://www.ussg.iu.edu/hypermail/linux/kernel/0512.3/0050.html might be
[11:45] <stgraber> pitti: just noticed something on the tracker, when you add upgrade "isos", set the version number to tribe-X
[11:45] <Hobbsee> pitti: neat
[11:45] <stgraber> pitti: as we don't "rebuild" them
[11:45] <pitti> ogra_test: thumbs up!
[11:45] <ogra_test> yeah
[11:46] <stgraber> ogra_test: you won't say that after having logged out :)
[11:46] <ogra> seb128, gsd was pretty slow on startup ... i had unthemed panels for nearly a minute (1Ghz system and 256M ram here)
[11:46] <Hobbsee> ogra_test: yay!
[11:46] <ogra> stgraber, still testing the server desktop
[11:46] <stgraber> k
[11:48] <ogra> stgraber, can you remember if there was one or two ssh processes left ?
[11:48] <ogra> i suspect scottie kills only the one running *inside* the tunnel,  but not the tunnel itself
[11:48] <stgraber> ogra: one
[11:48] <ogra> ah, thats likely it then
[11:48] <stgraber> ogra: well, after logout I mean
[11:49] <ogra> i guess we need a .pid file for the tunnel rpocess or so
[12:00] <Hobbsee> pitti: had any htoughts about who to persuade to write the release notes for ubuntu?  kubuntu looks to be done, or almost done
[12:01] <pitti> Hobbsee: I'll asked seb for the gnome bits, and I'll steal OO.o from Kubuntu and will fill in apparmor
[12:01] <Hobbsee> pitti: cool
[12:03] <Hobbsee> hmmm.  we should edit that kubuntu release notes ehre
[12:03] <seb128> I'm not good at making those look nice, etc
[12:03] <Hobbsee> seb128: that's why you try delegating first :)
[12:03] <stgraber> Testing seems to be going fine, just some Xubuntu testing still required (alternate) and maybe feisty->gutsy upgrade too
[12:04] <seb128> Hobbsee: especially when you are not native english speaker :p
[12:04] <Hobbsee> seb128: i can proofread, if that helps.
[12:04] <seb128> would be nice
[12:04] <seb128> I don't intend to write pages anyway
[12:04] <seb128> I'll let you know when I've edited them, thanks
[12:05] <pitti> seb128: I'll do some screenshots for fusa and tracker, ok?
[12:05] <seb128> pitti: great, thanks
[12:08] <ogra> seb128, trying to switch users on a thin client i get an error i dont have access to ~/.Xauthority (which is true since this is stored on the client)
[12:09] <ogra> seb128, the prob is that the error doesnt close on click
[12:09] <ogra> ah, using metacity works, but the close button doesnt
[12:10] <norsetto> ogra: have you received my email about amachu (new contributor)?
[12:10] <ogra> norsetto, yes, sorry i was very busy the last two days
[12:11] <norsetto> ogra: of course; just let me know when you have time; thx
[12:33] <pitti> seb128, Hobbsee: I pimped https://wiki.ubuntu.com/GutsyGibbon/Tribe4 a bit, WDYT?
[12:34] <seb128> pitti: very good ;)
[12:34] <stgraber> pitti: add drop-shadow to the screenshots :)
[12:35] <fabbione> oh feh.. pitti: RHCS is not critical for Tribe
[12:35] <fabbione> leave it queued for after
[12:35] <pitti> fabbione: RHCS?
[12:35] <thom> red hat cluster suite
[12:35] <fabbione> yeah i just uploaded redhat-cluster-suite
[12:38] <pitti> fabbione: ah, that's fine
[12:41] <asac> pitti: test install failed hard :(
[12:41] <Hobbsee> seb128: pitti stealing the wiki lock.
[12:41] <pitti> asac: ?
[12:41] <_StefanS_> hi there
[12:41] <asac> in the mids of installation ... i have no idea how i can get dmesg out of busy box to somehwere
[12:41] <pitti> Hobbsee: still heavily editing, please give me some minutes
[12:41] <asac> pitti: i used auto-resize option
[12:41] <_StefanS_> anyone know why subpixel hinting is not compiled in for gutsy ?
[12:41] <Hobbsee> pitti: ahhh, okay.
[12:41] <pitti> asac: hm, I tested that, too
[12:42] <asac> pitti: and see things like "attempt to access beyond end of device" ...
[12:42] <Hobbsee> pitti: was just going to add the ooo stuff in, but you can do that yourself if you've got teh wiki lock
[12:42] <asac> pitti: then falling back to ext2
[12:42] <pitti> Hobbsee: already done
[12:42] <asac> pitti: after install of xorg ... i suddenly got a corrupted curses screen: e.g. just stripes et al
[12:42] <asac> pitti: any idea how i can get the dmesg from busybox somewhere?
[12:43] <Hobbsee> pitti: even better.  wasnt there when i last refreshed :)
[12:43] <pitti> asac: not really :/ photo camera?
[12:43] <asac> hmmm not avail
[12:43] <asac> i have wget :) ... lets see if can pastebin that ;)
[12:45] <pitti> Hobbsee, seb128: https://wiki.ubuntu.com/GutsyGibbon/Tribe4 pimped further; I'll collect some caveats now and fill them in, rest is more or less complete, I think; review and fixing my Denglish highly appreciated
[12:46] <_StefanS_> bryce: ping ?
[12:46] <Hobbsee> pitti: seb128 OK, stealing the wiki lock, to fix the Denglish
[12:46] <seb128> Hobbsee: thanks ;)
[12:46] <seb128> pitti: looks good to me
[12:46] <_StefanS_> bryce: any idea why subpixel hinting is not compiled into gutsy for freetype ?
[12:47] <pitti> Hobbsee: can you please download tribe s/3/4/ while you are at it?
[12:47] <Hobbsee> yep
[12:49] <pitti> hm, the Xubuntu page doesn't look very good, I think I won't link to it
[12:49] <Hobbsee> pitti: okay, i see what you mean about the Denglish :P
[12:50] <pitti> Hobbsee: some Frenglish amongst it, too
[12:50] <pitti> Hobbsee: I'll look at the diff later to learn :)
[12:52] <Hobbsee> :)
[12:53] <Hobbsee> pitti: unsure why the deskbar applet is mentioned separately, and also in the tracker section.
[12:55] <pitti> Hobbsee: deskbar is not only for tracker searches
[12:55] <Hobbsee> pitti: true
[12:59] <mvo> Hobbsee: should we mention the improved apport reports on package fialures? I guess its too much of a non-desktop feature
[12:59] <pitti> mvo: I guess so; maybe we should not advertise things that make crashes look better too much :-)
[01:00] <Hobbsee> mvo: ask pitti.  i'm not around for this tribe :)
[01:00] <pitti> mvo: but if you want to write a stanza about it, go ahead please
[01:00] <mvo> pitti: not really, was just curious :)
[01:01] <thom> s/moving very quickly/hurrying/
[01:02] <thom> Gnome -> 'as always' > 'as usual'
[01:02] <pitti> Hobbsee: ^ you still have the wiki lock?
[01:03] <thom> Fast User: "Fast User Switching" rather than "Fast user switch applet" (dunno)
[01:03] <thom> Hobbsee: lemme know when you're done with the lock :)
[01:03] <pitti> thom: thanks
[01:03] <Hobbsee> thom: i shouldnt do, now
[01:06] <Hobbsee> thom: ah, good finds.   didnt notice them
[01:07] <asac> pitti: can't telll ... retry ... reuse previously created partition appears to work :/
[01:08] <pitti> Hobbsee, Riddell, ogra: at this stage I'm content with the Ubuntu CDs; I don't like the warts, but it installs at least, and it's too late for new images; what's the "yay or nay" status for kubuntu/edubuntu?
[01:08] <Hobbsee> pitti: yay for kubuntu - we would have had to respin them hours ago to do otherwise.
[01:08] <pitti> asac: that sounds serious, though; that should be a tribe-5 bug; it doesn't seem to happen generally, though
[01:08] <ogra> just running a i386 ubiquity install here, seems good so far
[01:08] <asac> pitti: there was a difference
[01:08] <ogra> i386 server is definately ok
[01:08] <asac> pitti: i had no monitor connected so X resolution selection popped up
[01:09] <thom> you're far too excited about OOo :)
[01:09] <asac> pitti: this time i kept monitor connected to see what is going on
[01:09] <pitti> asac: that always happends to me on alternates, too
[01:09] <ogra> i would have liked an amd64 test though, but my hw is broken atm
[01:09] <asac> pitti: resolution selection?
[01:09] <asac> pitti: hmm
[01:09] <pitti> ogra: stgraber did quite a lot of tests yesterday with the previous image, but not sure about amd64
[01:09] <ogra> theer are no amd64 results on the isotracker
[01:10] <pitti> ogra, Hobbsee: ok, then I'll prepare the /releases/ ISO images now; we can still ditch them if needed, but if they are reaonable, I have it prepared already
[01:10] <ogra> oh, wait
[01:10] <ogra> desktop had amd64 tests
[01:10] <Riddell> pitti: yay indeed
[01:10] <Hobbsee> pitti: great :)
[01:11] <pitti> Riddell: I like the OOo screenshot on the kubuntu page :-P
[01:11] <pitti> Riddell: "Yes, it's blue again!!!"
[01:13] <Riddell> makes me happy :)  I should update the page with info on how to get oo working
[01:13] <Hobbsee> Riddell: i did it
[01:13] <pitti> Riddell: it's there already
[01:14] <pitti> thom: thanks, adding my caveat list then
[01:14] <Riddell> oh, so it is
[01:15] <thom> someone please check my english; I've been living in .nl for the last year and it's doing bad things to my grammar!
[01:16] <pitti> wiki updated
[01:16] <Hobbsee> thom: why'd you take the comma out?
[01:16] <Hobbsee>  * On Kubuntu installs Open Office currently does not start but hangs at the splash screen. To workaround this install openoffice.org-gnome. (https://bugs.launchpad.net/bugs/127944)
[01:16] <ubotu> Launchpad bug 127944 in openoffice.org "[gutsy]  Open Office applications don't start " [High,Confirmed] 
[01:17] <Hobbsee> thom: the LTSP change looks wrong too - it should be in.
[01:17] <stgraber> ogra,pitti: my test computer is i386 only
[01:17] <Fujitsu>  /win 16
[01:17] <Hobbsee> thom: did you want to fix, or me to fix?
[01:17] <pitti> stgraber: alright; then let's hope for the best :)
[01:17] <Fujitsu> Oops.
[01:18] <thom> Hobbsee: no way, definitely "on the order"
[01:18] <thom> hrm, maybe not
[01:18] <Hobbsee> thom: everyone says "in the order of" here.  maybe it's an australian-ism.
[01:18] <thom> meh :)
[01:18] <Fujitsu> I don't think it's an Australianism.
[01:18] <Hobbsee> but, i dont think so
[01:18] <ogra> stgraber, all fine, there was one amd64 -desktop test
[01:18] <ogra> i'm fine to take the risk for amd64 server
[01:19] <thom> what don't you like about the openoffice change? ", hanging" -> "but hangs" seems reasonable to me, but like i say :)
[01:19] <thom> Hobbsee: you're welcome to fix :)
[01:20] <Hobbsee> thom: just wondered why you dropped the comma
[01:20] <Hobbsee> thom: that was the only thing
[01:23] <Hobbsee> thom: fixed
[01:27] <thom> cools
[01:27] <thom> i suspect i was just on an anti-comma tear :)
[01:27] <Hobbsee> thom: hehe
[01:29] <ogra> seb128, we should probably add a casper script to disable fusa in the live session weems useless to me to have it there
[01:29] <ogra> *seems
[01:48] <pitti> Hobbsee, seb128, thom: are you happy with the Ubuntu wiki page now? if so, I'll move it over to www.ubuntu.com
[01:52] <seb128> ogra: for edubuntu you mean?
[01:52] <ogra> seb128, generally
[01:53] <ogra> seb128, what the use case to have it if there is only a live session user
[01:53] <seb128> ogra: I don't agree, it display who is logged, allow to right click to add users and is consistant
[01:53] <seb128> pitti: looks good to me
[01:55] <cjwatson> asac: you can extract log files using nc, or you can 'anna-install openssh-client-udeb' and use scp
[01:56] <ogra> or use the menu to do the latter ;)
[01:56] <asac> damn ... the state is gone already :(
[01:56] <asac> cjwatson: ^^
[01:58] <cjwatson> asac: ok, but for the future
[01:58] <asac> yes .. will remember anna-install
[02:10] <Admiral_Chicago> pitti: how long until the release?  I'm still working on the notes
[02:11] <pitti> hi Admiral_Chicago
[02:11] <Admiral_Chicago> morning
[02:11] <pitti> Admiral_Chicago: I think two or three hours, since I have to wait for our webmaster
[02:11] <Hobbsee> pitti: you might be lucky, and get it out before the distro team meeting
[02:11] <pitti> Hobbsee: that's my hope
[02:12] <Admiral_Chicago> pitti: okay the Xubuntu notes aren't done yet, unfortunately.
[02:12] <pitti> Admiral_Chicago: I saw; if you want to do some, that would be great
[02:13] <doko> pitti: please accept php5 (lpia fix), if this doesn't show up on the CD
[02:13] <Admiral_Chicago> i'll see what I can do right now, have to go to work in a few minutes, thats been takig up all my time :(
[02:23] <stgraber> ouch, just noticed that the tracker now takes >1 hour to update it's LP bug DB, I'll put caching of bug status (not refresh not shown bug) high priority
[02:46] <pitti> cjwatson, evand: I filed bug 131294 (for tribe-5 only); I assigned it to evand for now, please trade it amongst the two of you accordingly :)
[02:46] <ubotu> Launchpad bug 131294 in ubiquity "does not install language packs for the target language" [High,New]  https://launchpad.net/bugs/131294
[02:47] <cjwatson> pitti: evand is welcome to it :-)
[02:48] <pitti> heh, declining the honour? :)
[02:48] <Hobbsee> hehe, delegation at it's finest
[02:49] <Hobbsee> pitti: do my assignment, please.  it'll be good for you.
[02:50] <pitti> Hobbsee: you are talking to a shell script (~/bin/pitti.sh); I am not capable of such things
[02:50] <Hobbsee> pitti: awww....darn.
[02:51] <Hobbsee> $ for x in lots; do cp ~/bin/pitti.sh ~/bin/pitti-$x; done.
[02:51] <Hobbsee> pitti: if you're a shell script, then we can clone you!
[02:52] <Hobbsee> excellent!
[02:52] <Spads> set -e
[02:52] <Hobbsee> Spads: that too.  it was pseudocode :P
[02:52] <Spads> Hobbsee: no, not for you
[02:52] <Spads> Hobbsee: for pitti
[02:52] <Kmos> pitti: http://cdimage.ubuntu.com/dvd -> this one isn't building daily
[02:52] <Spads> Hobbsee: make him less tolerant of error!
[02:52] <Hobbsee> Spads: ahhh....
[02:53] <pitti> Kmos: I know, they are horribly overflown, and I don't have an idea ATM how to fix that
[02:53] <Kmos> http://cdimage.ubuntu.com/dvd/20070804/gutsy-dvd-amd64.OVERSIZED
[02:53] <Kmos> :)
[02:53] <Kmos> hehe
[02:54] <Zic> Tribe 4 is online
[02:54] <Zic> :] 
[02:55] <Kmos> pitti: bug 130777
[02:55] <ubotu> Launchpad bug 130777 in Ubuntu "Current Gutsy DVD build ISO is too large" [Undecided,Triaged]  https://launchpad.net/bugs/130777
[02:55] <Kmos> pitti: how about to delete the current directory, so anyone will download it..
[02:57] <cjwatson> I'd rather have a clearer note on the page that the image is oversized, so that we don't impede rsyncability for developers keeping up
[02:57] <cjwatson> anyone in the ubuntu-cdimage team has code access and is in a position to implement
[02:57] <cjwatson> that
[02:57] <cjwatson> Kmos: DVDs aren't supposed to build daily, either.
[02:57] <cjwatson> that's deliberate
[02:57] <cjwatson> would be too much mirror churn
[02:59] <Kmos> cjwatson: I've already explained in the man of the bug that he must check of .OVERSIZED, so he don't have 4 hours of download for nothing
[03:00] <cjwatson> I saw, but it would also be possible to explain that in prose in the automatically generated HTML
[03:00] <cjwatson> (I'm not going to just edit it on the fly every time; the code should be changed)
[03:01] <Kmos> yeah, I agree
[03:01] <cjwatson> maybe we should create a dvd seed, though I was hoping not to have to do that
[03:02] <calc> cjwatson: oh ok
[03:03] <calc> pitti: i think i identified the library that allows it to work at all, not the one that needs to be fixed
[03:03] <pitti> calc: ooh
[03:03] <cjwatson> doko: would you object to dropping gcc-4.2-source from supported (and thereby from the DVD)? it will stay in main due to build-dependencies
[03:03] <pitti> calc: we currently recommend to install openoffice.org-gnome as a workaround
[03:03] <cjwatson> and will save 50MB of the 170MB or so we need to trim
[03:03] <pitti> calc: but that itself isn't needed; which package is it then?
[03:04] <cjwatson> calc: "cjwatson: oh ok" => sorry, context?
[03:04] <calc> pitti: apparently that helps for ubuntu but not kubuntu, i still need to do more investigation into it
[03:04] <pitti> calc: ok; well, I'll just leave the ooo-gnome workaround in the Kubuntu release notes then
[03:04] <calc> pitti: from what i read on a bug report its hangs due to not properly initializing gtk
[03:04] <cjwatson> it's awfully tempting to drop linux-source from the DVD but I think that's useful to have there
[03:05] <calc> cjwatson: ah about what you said about ingo and noatime
[03:05] <cjwatson> oh, right
[03:05] <cjwatson> calc: (this morning, I committed support to d-i upstream for relatime)
[03:05] <calc> cjwatson: ok
[03:05] <cjwatson> still need to do support for having default mount options in a reasonable way, though
[03:06] <pitti> ogra, Riddell: did you find any new OMGTSIF bugs? speak now or never (about to sync-mirrors, for link/bittorrent testing etc)
[03:06] <Riddell> pitti: nope, go ahead and sync
[03:08] <cjwatson> pitti: I think bzip2ing texlive-doc and texlive-lang would help
[03:08] <cjwatson> maybe even texlive-base
[03:08] <cjwatson> I'm just looking into the numbers
[03:09] <pitti> cjwatson: would mean a (trivial) debian delta, but indeed
[03:09] <cjwatson> pitti: yes, seems like it might be worth it
[03:09] <cjwatson> no intrinsic reason Debian couldn't take it
[03:13] <evand> pitti: I'll take it (not that I have a choice :) )
[03:14] <pitti> evand: thanks
[03:14] <cjwatson> hmm, not obvious that it saves a whole lot from texlive-doc
[03:16] <cjwatson> damn, bzip2ing texlive-* seems to save maybe a megabyte at most from the largest ones. Probably not worth it :(
[03:16] <ScottK> Hobbsee: If you have an opinion, please express it.
[03:18] <cjwatson> hmm, I wonder if I should make germinate show original-maintainer rather than maintainer
[03:18] <cjwatson> seems like it would be more informative
[03:27] <pitti> hi mathiaz
[03:29] <mathiaz> hi pitti
[03:29] <mathiaz> pitti: how things are going with the isos ?
[03:29] <pitti> mathiaz: last stages of preparation now
[03:29] <iwj> pitti: I see you marked my `fsck on first boot' bug 131201 as a dupe of bug 63175 but I don't think that's right.
[03:29] <pitti> mathiaz: a lot of warts on the desktop, but too late now
[03:30] <mathiaz> pitti: server seems all good
[03:30] <iwj> The original submitter of 63175 is getting fsck on _every_ boot.
[03:30] <ubotu> Launchpad bug 131201 in debian-installer "always fscks first boot after install (dup-of: 63175)" [Undecided,Confirmed]  https://launchpad.net/bugs/131201
[03:30] <ubotu> Launchpad bug 63175 in e2fsprogs "fsck on every (re)boot" [Medium,Incomplete]  https://launchpad.net/bugs/63175
[03:30] <iwj> It's possible the cause is the same but I doubt it.
[03:30] <pitti> iwj: hm, the bug trail seemed to indicate that this was related (clock skew)
[03:30] <iwj> I think it may well be related, yes.
[03:30] <pitti> iwj: but I'm ok with unduplicating it; I didn't look into it very deeply
[03:30] <dendrobates> pitti, mathiaz: I was unable to test the sparc iso.  My ultra 10 needs a little loving care first.  I'll test this afternoon.
[03:31] <pitti> dendrobates: alright; we'll just push it out and hope then :)
[03:31] <iwj> I don't think we know how to reproduce 63175.
[03:33] <ogra> iwj, thst bugs exists since tribe 1 ... there must me another one that matches better iirc
[03:33] <iwj> ogra: Right.
[03:33] <ogra> its related to UTC/non UTC setup ....
[03:34] <doko> cjwatson: np, just needed to build gcj-4.2
[03:34] <dobey> hola
[03:36] <cjwatson> fsck> I would love somebody who can reproduce it to fix it
[03:37] <iwj> cjwatson: 63175 or the first-boot one ?
[03:37] <cjwatson> iwj: either, frankly :-)
[03:37] <iwj> :-).
[03:37] <iwj> 63175 is too hard I think.
[03:37] <cjwatson> but I meant the first-boot one
[03:37] <iwj> But I can reproduce the latter.
[03:37] <iwj> I'll look at it after FF.  I take it we're not in a hurry.
[03:37] <cjwatson> I am available for advice if anyone can describe the problem within d-i accurately and is just trying to figure out the right fix
[03:38] <cjwatson> hurry> shouldn't think so
[03:38] <iwj> cjwatson: Right.
[03:39] <iwj> I've assigned the bug to myself and milestoned it for tribe 6.  Is that about right ?
[03:39] <cjwatson> sounds lovely
[03:40] <cjwatson> (tribe 5 is after FF too, but tribe 6 might be more realistic since I suspect it may take a little while to sort out)
[03:40] <cjwatson> iwj: FWIW /var/log/installer/cdebconf/questions.dat should be enough information for accurate reproduction by somebody who knows how to read it; alternatively an install with DEBCONF_DEBUG=developer in the kernel boot parameters
[03:40] <cjwatson> (and /var/log/installer/syslog from the latter)
[03:41] <dobey> where does the standard .config for building an ubuntu kernel come from? linux-source doesn't seem to include it, and i'm not sure it has the "ubuntu patches" in it either
[03:41] <iwj> cjwatson: Right.  I'll C&P that into the bug so I'll know what to do later ..
[03:42] <cjwatson> dobey: apt-get source linux-source-2.6.22 (or appropriate other version)
[03:43] <cjwatson> it's in debian/config/ or something like that, and the patches are applied
[03:43] <dobey> ah
[03:44] <ogra> iwj, i think we already hunted it down to something like: partitioner runs mkfs and puts timestamp into superblock, timezone gets configured and changes the clock backwards (UTC setting), superblock is in the future on first boot and triggers fsck
[03:46] <iwj> Yes.
[03:46] <iwj> The no. of days is 2^32s - 86400 + 3600 or some such.
[03:49] <dobey> hrmm
[03:50] <dobey> this one ps3 patch seems to already be applied...
[03:52] <cjwatson> dobey: certainly a number of ps3 patches are
[03:53] <dobey> cjwatson: i wonder why my sixaxis doesn't work over usb though, given the appropriate patch seems to be applied already
[03:53] <dobey> :-/
[03:54] <cjwatson> dobey: worth asking #ubuntu-kernel; the relevant people should have woken up / be waking up shortly
[03:55] <mjg59> iwj: Can you see a downside to just altering the fsck code so it ignores dates in the future?
[03:56] <dobey> interesting
[03:57] <iwj> mjg59: If the time check interval thingum is set to `don't' then yes, it should ignore them.
[03:58] <iwj> Otherwise it's not clear.
[03:59] <mjg59> iwj: My recollection is that older versions of fsck did signed arithmetic here
[03:59] <sladen> isn't it just a (signed) issue?
[03:59] <mjg59> Someone found the old code at some point
[03:59] <iwj> I worry though about fixing this particular symptom like that.  There might be other things that go wrong as a result of clock skew between install and first boot.
[04:00] <iwj> signed arithmetic> That means if you once fsck with a stupid clock, you might never fsck again.
[04:00] <mjg59> Hm. That might be the reason for the change
[04:01] <mjg59> On the other hand, if you change the hardware time, it's not obvious that that should result in a fsck
[04:01] <sladen> so have a "uninitialised" state (eg. 0x...fffff) that is treated as such
[04:01] <sladen> set that during that one-time following creation
[04:01] <mjg59> The obvious way around this would seem to be to write the current time if the existing time is in the future
[04:02] <pitti> have fun
[04:02] <StevenK> pitti: Yay! Congrats
[04:02] <pitti> what a messy release again
[04:02] <ogra> milestone ...
[04:02] <ogra> :)
[04:02] <Hobbsee> yay, pitti!
[04:03] <pitti> at least I learned a whole lot about the guts of Soyuz :) (more than I wanted to, I guess)
[04:11] <Hobbsee> what do we do about bugs such as https://launchpad.net/bugs/131311 ?
[04:11] <ubotu> Launchpad bug 131311 in Ubuntu "tribe cd images not being produced for powerpc" [Undecided,New] 
[04:12] <StevenK> I think there was no point making a live CD for powerpc prior to OO.o 2.3, since it didn't build on powerpc
[04:13] <StevenK> Hrm. No, that doesn't seem to be it, I'll shush now.
[04:13] <asac> bryce: yt?
[04:15] <Hobbsee> asac: btw, any plans to update enigmail?
[04:15] <asac> Hobbsee: is there a wishlist bug ;) ?
[04:15] <asac> anyway, yes.
[04:15] <Hobbsee> asac: dunno.  havnet looked.
[04:16] <Hobbsee> asac: i looked at upgrading it, and went "ewww, that looks decidedly not-fun"
[04:16] <asac> Hobbsee: its just dropping the enigmail tarball in the archives dir
[04:16] <asac> bumping changelog ... then spin
[04:16] <Hobbsee> asac: i wasnt sure if there was more than that
[04:16] <asac> usually not
[04:17] <asac> if it breaks, then come back :)
[04:17] <asac> just be sure that you keep the mozsdk tarball in orig as well
[04:17] <asac> just upgrade the enigmail one
[04:20] <Hobbsee> right
[04:28] <bddebian> Heya
[04:49] <pitti> StevenK: right
[04:50] <pitti> Hobbsee, StevenK: OO.o has never been built on ppc so far
[04:50] <pitti> Hobbsee, StevenK: i actually did attempt to build tribe4 images for ppc, but unfortunately they are uninstallable
[04:50] <pitti> which is actually my bug (restricted-manager), already fixed in bzr head
[04:50] <pitti> so we'll get dailies in the next days
[04:52] <pitti> jamiemcc: it's sooo great to have an upstream who directly works on LP bugs \o/
[04:52] <jamiemcc> pitti: no  problem - you lot have my undivided attention
[04:58] <bryce> morning
[04:58] <pitti> hey bryce
[05:04] <pitti> lamont`: just use requestsync :)
[05:04] <geser> preferable the version from gutsy
[05:04] <lamont`> pitti: I even have that
[05:05] <lamont`> pitti: whatever version >= 1:1.0.0-7 will prevent you and fabbione from hurting me
[05:06] <lamont`> er, -8
[05:06] <fabbione> lamont`: if it's only because of portmap, i won't hurt you too badly :)
[05:06] <lamont`> fabbione: nfs mounting was dropped by upstream util-linux...  nfs-utils now delivers mount.nfs
[05:07] <fabbione> oh i see
[05:07] <fabbione> well i will put the package on hold
[05:07] <lamont`> and mount doesn't depend: nfs-utils since 99% of the world doesn't care about nfs, and grandma will never figure out how to remove portmapper later.
[05:08] <fabbione> yeah i know i am that 1% special
[05:08] <fabbione> that makes me feel so gooood
[05:08] <lamont`> preinst has logic that says: if nfs-common (any  version) is installed, and proc is mounted and there are nfs mounts in /proc/mounts, and /sbin/mount.nfs doesn't exist, then bitch and die
[05:08] <fabbione> gotcha
[05:08] <lamont`> the further justification is that it is reasonable to expect that anyone running nfs has sufficient clue to deal with "upgrade nfs-common first, or die" error message.
[05:10] <lamont`> pitti: looks  like he actually got done fixing mount-related bugs in 1:1.1.0-14 :)
[05:10] <fabbione> lamont`: yeah i guess i fall into the "enough clue" by being part of that 1% :P
[05:10] <lamont`> fabbione: note that I didn not say "old farts" even once.
[05:11] <fabbione> lamont`: ROFL
[05:11] <fabbione> but then.. i already use postfix...
[05:11] <norsetto> hey, mind your language
[05:11] <norsetto> old is a bad word :-)
[05:19] <alex-weej> cjwatson: got a second mate?
[05:20] <cjwatson> alex-weej: not really, distro team meeting in progress
[05:20] <cjwatson> alex-weej: you can let me know what's up and I can get back to you after the meeting's over
[05:21] <alex-weej> cjwatson: ok thanks, i just want to be sure of how to get Tribe 4 installed so i can test it without wasting an hour installing and then have the bootloader installation failed
[05:22] <cjwatson> as I say, use the device name (/dev/sdd) without trying to translate to grub-speak yourself
[05:22] <alex-weej> ok i'll give it a go then, cheers!
[05:40] <ScottK> pitti: I really like apport saying "The problem cannot be reported.  This is not a genuine Ubuntu package."  Very good that it checks.  Thank you.
[05:41] <cjwatson> IMO it should just go away rather than telling you that though
[05:41] <cjwatson> as in, no UI until it's confirmed that it's actually a package bug
[05:41] <cjwatson> (if possible)
[05:42] <pitti> ScottK: :)
[05:43] <pitti> cjwatson: bit tricky, since we already have UI to restart the program and inform about the crash
[05:43] <pitti> and at that time we don't know yet whether it's a genuine package
[05:43] <pitti> (it is expensive to find out, and I don't want to do it unless the user is interested in reporting)
[05:43] <ScottK> Would that be the preferred approach then for a developer who wants a retrace on a crash they are working on?
[05:43] <pitti> but I do acknowledge that this could be better
[05:54] <Skiessi> why gutsy has still SDL 1.2.11? why not 1.2.12?
[05:54] <cjwatson> Debian hasn't packaged 1.2.12 yet, so we probably simply didn't notice
[05:54] <keescook> Skiessi: looks like debian hasn't packaged 1.2.12 yet: http://packages.qa.debian.org/libs/libsdl1.2.html
[05:54] <cjwatson> what is interesting in 1.2.12?
[05:55] <pitti> ScottK: I have some ideas about that, liek providing an 'debug this right here' know if apport-retrace and the ddeb repo are available, and such
[05:55] <cjwatson> (i.e. why should we take the extra effort to work in advance of Debian in this case?)
[05:55] <Skiessi> they have fixed some things in mouse support and my cursor has got stuck after quitting some sdl games
[05:56] <ScottK> pitti: I think that would be even better.  That way LP doesn't get cluttered.
[05:56] <Skiessi> so I had to always restart X server
[05:56] <Yagisan> Skiessi, code around the known mouse issues - restarting sdl apps usually fixes the mouse ...
[05:57] <mathiaz> keescook: So what's the plan to update apparmor kernel module ?
[05:57] <keescook> pitti: I looked at the apparmor profile for CUPS.  why the need for shadow?  Seems like the 'authentication' abstraction would be useful there?
[05:57] <keescook> mathiaz: it's up to the kernel team
[05:57] <mathiaz> keescook: I've updated the userspace part, but cannot test it as it needs the latest version.
[05:57] <keescook> mathiaz: we're waiting for them to take the patches
[05:57] <keescook> we discussed that it would be after t4
[05:57] <mathiaz> keescook: Have you sent the patch ?
[05:58] <keescook> mathiaz: my understanding is that kyle has it already
[05:58] <shirish> pitti: you there m8?
[05:58] <mathiaz> kylem: do you have the new apparmor kernel module ?
[05:59] <kylem> did you send it to me?
[05:59] <pitti> keescook: yeah, probably; it uses <shadow.h> and getspnam() etc
[05:59] <tkamppeter> pitti, another important thing where I am in favour of s-c-p are the bugs of g-c-m.
[05:59] <keescook> kylem: I thought you said you had the 10_3 branch from suse?
[05:59] <pitti> tkamppeter: I agree
[05:59] <pitti> shirish: yeah, barely
[06:00] <keescook> kylem: I had sent you the links, and you and BenC looked at them for a bit.
[06:00] <kylem> oh, right.
[06:00] <shirish> pitti: do you have some tests, I am getting 502 bad gateway whenever i try to push a 12 MB crash file through apport.
[06:00] <pitti> shirish: sounds like a Malone problem
[06:01] <shirish> pitti: could it be? How do I know that?
[06:01] <tkamppeter> pitti, g-c-m was not maintained upstream, so no one worked on the bugs. s-c-p is maintained very actively, and used on the most widespread distro, so bugs get fixed there.
[06:01] <pitti> keescook: I'm happy to use the auth abstraction, though
[06:02] <keescook> pitti: okay, I may tweak this a bit.  I think we should likely add an "avahi-client" abstraction too
[06:02] <pitti> keescook: cupsd checks users and passwords (lpadmin members) on the web interface
[06:02] <pitti> tkamppeter: would you have some time for some easy UI tweaks? like we discussed some days ago?
[06:03] <tkamppeter> pitti, a tool full of bugs is less user-friendly than a tool which asks perhaps a question too much.
[06:03] <pitti> tkamppeter: I'm also leaning towards using s-c-p, if we can make it a bit easier, it'll be great
[06:03] <pitti> tkamppeter: *nod*
[06:04] <tkamppeter> pitti, I have also forwarded your UI suggestions to Tim Waugh, perhaps he will also do something on them.
[06:04] <pitti> keescook: btw, how evil it is to do "/usr/lib/** mr,"?
[06:04] <pitti> keescook: I guess that pretty much circumvents the ixr limitations on /usr/bin/, right?
[06:04] <pitti> keescook: I put that in to get a profile ready at that day, but I guess it should be tighter
[06:04] <pitti> tkamppeter: yay, thanks
[06:06] <mathiaz> pitti: it should be included with the base abstraction.
[06:06] <pitti> mathiaz: no, that only has /usr/lib/**  are
[06:07] <pitti> mathiaz: oh, sorry, /usr/lib/*.so*                 mr, too
[06:07] <pitti> 'are'? I didn't type that, that was 'r'
[06:07] <pitti> silly xchat :)
[06:08] <mathiaz> pitti: access to libraries should be covered in the base abstraction.
[06:09] <pitti> mathiaz: what I meant is, if you allow loading and using *all* libraries, then it does not make sense to restrict execution privs to /usr/bin/, or does it?
[06:09] <pitti> mathiaz: since in all those libraries there's certainly some code you could abuse
[06:09] <pitti> mathiaz: hm, no idea why I added "/usr/lib/** mr" explicitly; I used it with the profile builder and it barked on me
[06:10] <pitti> mathiaz: I'll fine-tune that, I guess
[06:10] <mathiaz> pitti: there is also /lib/** rm
[06:10] <mathiaz> pitti: there is /lib/lib*.so*                  mr
[06:10] <mathiaz> pitti: in the base abstractions
[06:11] <alex-weej> cjwatson: it almost worked. except that if i attempt to boot from the external disk, grub can't bootstrap because it doesn't recognise the file system (Error 17). if i boot from my working grub and use a shell, it shows the filesystem as unknown. the file system is definitely ext3 and i can mount it here fine.
[06:12] <cjwatson> alex-weej: sorry, I can't help, but I hope somebody else can
[06:12] <alex-weej> ok thanks
[06:13] <alex-weej> i will put it on launchpad
[06:13] <alex-weej> does anyone know where the installer logs go?
[06:13] <cjwatson> /var/log/installer/
[06:15] <cjwatson> pitti: thanks for the vmware monitor size tip, btw
[06:15] <pitti> cjwatson: heh, I was annoyed by it and I just happened to discover that it's configurable, no problem :)
[06:16] <cjwatson> Riddell: turns out the failure to do a CD check is directly due to there being no suitable usplash theme
[06:16] <cjwatson> I managed to strace it, and it does open("/dev/.initramfs/usplash_fifo") => ENXIO and bails out there
[06:17] <keescook> pitti: /usr/lib/** mr, seems pretty open, but probably not hugely harmful.  an attacker already can read /etc/shadow, for example
[06:17] <cjwatson> might be able to make it fall back to text-only
[06:18] <pitti> keescook: I agree
[06:18] <pitti> keescook: what I meant is, with abstractions/base it does to make too much sense to focus on restricting /usr/bin executability, right?
[06:19] <Riddell> cjwatson: oh, curious
[06:19] <keescook> pitti: well, just because you can mmap it, doesn't mean you can exec it.  :)
[06:19] <keescook> or, rather, it raises the bar a bit
[06:20] <pitti> keescook: but /usr/lib/** has certainly a lot of code to exploit/use
[06:20] <pitti> keescook: yeah, it probably defeats standard exploits pretty well
[06:20] <pitti> keescook: I'm thinking on the academic level now
[06:20] <pitti> i. e. 'someone really means it'
[06:30] <pitti> keescook: so I'll clean up the profile for abstractions/auth and the reduntant library stuff; did you see anything that's badly wrong or horrifying?
[06:49] <ubotu> Launchpad bug 128744 in soyuz "spurious titlecase in PPA indices" [High,In progress]  https://launchpad.net/bugs/128744
[06:49] <bryce> mvo, did you want to chat?
[06:49] <mvo> bryce: I asked in ubuntu-x :)
[06:49] <bryce> ahh
[06:51] <alex-weej> http://hoegsberg.blogspot.com/2007/08/redirected-direct-rendering.html
[06:51] <alex-weej> we surely need this for gutsy if we're going to ship compiz on by default
[06:51] <thom> alex-weej: for gutsy? dream on
[06:51] <thom> it's not even upstream yet
[06:51] <alex-weej> but we're still going to ship compiz on?
[06:53] <mjg59> alex-weej: It's only implemented for Intel, it's implemented for a branch of the intel driver we don't use, it's implemented for a branch of drm we don't use and it requires TTM support (which is unstable and which we don't use)
[06:53] <mjg59> So that's certainly not going in gutsy
[06:53] <alex-weej> .'. no default compiz
[06:54] <pitti> LaserJock: \o/
[06:54] <alex-weej> i don't wanna imagine the stick that Gutsy gets when average Joe realises that he can't have "Vista-style" 3D whack while being able to use Google Earth
[06:55] <mvo> alex-weej: that and video are currently the two big showstoppers for compiz
[06:55] <alex-weej> what's wrong with video?
[06:55] <pygi> alex-weej, I can imagine. You are getting it all wrong anyway
[06:55] <pygi> whatsoever, I have and won't go into further discussion
[06:56] <mvo> you can not have composited video currenly (only with a specially patched mplayer)
[06:56] <alex-weej> mvo: hm, video works for me with R350 on Gutsy, as long as i don't use Xv (naturally)
[06:57] <mvo> alex-weej: right, that is a workaround of course, but non-Xv is not that great performancewise
[06:57] <alex-weej> what's the solution?
[06:57] <mjg59> Use textured video rather than overlay
[06:57] <mjg59> (= wait for the code to be written)
[06:57] <alex-weej> GL?
[06:58] <mvo> Riddell: https://bugs.launchpad.net/ubuntu/+source/update-manager/+bug/131352 looks like something releated to kde?
[06:58] <ubotu> Launchpad bug 131352 in update-manager "upgrade tool crashed in Gutsy Tribe 3" [Undecided,New] 
[06:58] <alex-weej> is that what GL output is?
[06:58] <mjg59> But yeah. Right now compiz causes serious regressions.
[06:58] <alex-weej> so it's looking like Compiz won't be on by default after all?
[06:58] <mjg59> Well, it still has exactly the same issues that made us punt it last time
[06:59] <alex-weej> PUNT
[06:59] <alex-weej> whenever i read that word i think of Teen Girl Squad. Excuse me.
[07:00] <Riddell> mvo: yes, assign to me if you want
[07:02] <mvo> Riddell: do you know already what the cause is?
[07:02] <mvo> Riddell: thanks, assigned it to you now
[07:03] <Riddell> mvo: no idea, but we have new versions of pyqt and pykde in recently
[07:04] <mvo> Riddell: bug #130321 is something that would be good to get solved for gutsy too, the forked dpkgpm causes more and more pain now
[07:04] <ubotu> Launchpad bug 130321 in adept "apport integration for package failure does not work with adept" [Undecided,New]  https://launchpad.net/bugs/130321
[07:04] <mvo> Riddell: essentially adept lacks automatic-dependency information, dpkg-log support and apport integration because of it
[07:05] <mvo> Riddell: any hackers you can summon on it? if not, I will see if I can give it a go, but I will take twice the time of a kde hacker :/
[07:44] <superm1> join #ubuntu-chicago
[08:02] <mjg59> jamiemcc: Doing a git pull on a small source tree results in tracker spending really quite a new time updating the index
[08:02] <mjg59> s/new/long/
[08:03] <mjg59> It's now spent over a minute writing to the depot files
[08:04] <mjg59> jamiemcc: strace shows it doing a lot of seeking, reading and writing
[08:04] <jamiemcc> mjg59: thats correct
[08:04] <jamiemcc> but it should be quicker
[08:04] <mjg59> 3 minutes in total
[08:04] <jamiemcc> kernel io performance on gutsy sux atm o n my hardware
[08:05] <jamiemcc> on feisty we can update 100 source files a second
[08:05] <mjg59> jamiemcc: The quantity of seeking its doing would result in it being slow anyway
[08:05] <mjg59> Is the file format really optimal?
[08:05] <jamiemcc> yes its a differential indexer
[08:05] <jamiemcc> we need to fetch the existing contents and compare with changed
[08:05] <jamiemcc> and then update difference to index
[08:06] <jamiemcc> mjg59: you can use our crawl directory settings to only index that stuff at startup
[08:06] <mjg59> Ok. I'm not especially happy about an update of 30 files resulting in tens of thousands of seek.
[08:07] <mjg59> That's going to be slow, especially on laptops
[08:07] <jamiemcc> 30 files means about 30 seeks
[08:07] <jamiemcc> + 30 for fetching original text
[08:07] <jamiemcc> should be 60 in total
[08:07] <mjg59> jamiemcc: That's not what I'm seeing
[08:08] <jamiemcc> you must have a huge db then
[08:08] <jamiemcc> and not much ram
[08:08] <mjg59> I have 768MB of RAM
[08:08] <mjg59> And a 1.3GB database
[08:08] <jamiemcc> how big is db files in ~/.cache
[08:08] <mjg59> 1.1G    /home/mjg59/.cache/tracker
[08:09] <mjg59> jamiemcc: http://paste.ubuntu-nl.org/33161/
[08:09] <mjg59> That's a snippet - imagine hugely more than that
[08:09] <jamiemcc> ok
[08:10] <jamiemcc> should not take a minute to do that
[08:10] <jamiemcc> a few secs is more normal
[08:10] <mjg59> Took 3
[08:10] <jamiemcc> 3 minutes?
[08:10] <mjg59> Yes
[08:10] <jamiemcc> how big were the files?
[08:10] <mjg59> All the time taken was writing to the database, not reading the files
[08:11] <jamiemcc> we have fsync off so the disk scheduler should optimise that
[08:11] <jamiemcc> I bet kernel is not doing that
[08:11] <mjg59> You're repeatedly seeking and reading
[08:11] <mjg59> There's little the kernel can do to optimise that other than readahead
[08:12] <jamiemcc> the wrotes should be buffered
[08:12] <mjg59> Right, but the reading clearly can't be
[08:12] <mjg59> And it's doing as many reads as it is writes
[08:13] <jamiemcc> if thwe writes are not buffered it will really slow things down
[08:13] <jamiemcc> the seeks are all close together on disk so should be fast
[08:14] <jamiemcc> if buffered the writes will be applied long after the seeks
[08:15] <mjg59> If you're making several thousand small reads, it's going to be slow
[08:15] <highvoltage> what's the status of Gobuntu? is there any value in testing the current builds?
[08:15] <highvoltage> (or tribe 4, at least)
[08:15] <jamiemcc> mjg59: no more than a few secs
[08:15] <jamiemcc> alternate read/write is painfully slow
[08:15] <mjg59> jamiemcc: Hang on, I'll try to instrument this
[08:15] <highvoltage> (oops, I see there's just daily builds currently)
[08:16] <mjg59> jamiemcc: Remember that with a small read followed by another small read, you're having to wait for the disk to spin round to the right place again
[08:16] <jamiemcc> mjg59: can you graph disk reads/writes and see when they are applied?
[08:16] <jamiemcc> mjg59: we dont control reads - sqlite does that
[08:16] <mjg59> So on a 4800RPM disk, you're looking at only being able to make 4800 small reads a minute
[08:16] <mjg59> (worst case)
[08:17] <jamiemcc> 10 ms access time is average
[08:17] <jamiemcc> for reads close together buffer will cover some of them
[08:17] <jamiemcc> and head has less distance to cover
[08:17] <mjg59> 10ms is only 6000 reads a minute
[08:18] <jamiemcc> thats worst case
[08:18] <jamiemcc> if they are close together average seek will be smaller
[08:18] <mjg59> Whether any of it gets buffered depends on whether the reads are close to each other
[08:18] <jamiemcc> yes
[08:18] <mjg59> It's not seek time that's the killer here. It's rotational latency
[08:18] <jamiemcc> ok
[08:18] <jamiemcc> but not sure what I can do about it?
[08:18] <jamiemcc> all i can say is its fast if the writes are buffered properly
[08:18] <mjg59> Let me just get an idea about how many reads are being made
[08:21] <rulus> jamiemcc: I mailed you a Dutch translation for Tracker yesterday, did you receive the message?
[08:22] <jamiemcc> rulus: yes - I will apply soon thx
[08:22] <rulus> ok, thanks :)
[08:22] <rulus> it's not 100% complete, but it's a good start
[08:23] <jamiemcc> ok sorry for the delay applying - just be swamped with bugs atm
[08:23] <rulus> no problem
[08:26] <mjg59> jamiemcc: Ok, I'm benchmarking the creation of 399 files totalling 5.9M
[08:26] <mjg59> It's taken 3 minutes so far
[08:26] <jamiemcc> that sux
[08:27] <jamiemcc> are they text files?
[08:28] <mjg59> C files
[08:28] <jamiemcc> ok that should take less than 10 secs
[08:28] <mjg59> With a binary pack file
[08:28] <mjg59> So a single 2.8M binary
[08:29] <mjg59> Still running, so 6 minutes so far
[08:29] <Skiessi> could someone add tabbing to Nautilus? before or after gutsy
[08:29] <jamiemcc> mjg59: are other disk intensive processes slow?
[08:29] <mjg59> No
[08:30] <mjg59> Well, to the extent that this is a 1.8" hard drive, yes
[08:30] <jamiemcc> I mean ones that read/write a lot
[08:30] <mjg59> But otherwise, no - no slower than with older kernels
[08:30] <jamiemcc> are you in swap?
[08:30] <mjg59> There is swap in use, but I have plenty of free memory
[08:30] <jamiemcc> ok
[08:31] <mjg59> So running stuff isn't swapping
[08:31] <jamiemcc> ext3 ordered mode?
[08:31] <mjg59> All default mount options
[08:31] <jamiemcc> ok
[08:31] <mjg59> [    5.672000]  EXT3-fs: mounted filesystem with ordered data mode.
[08:32] <mjg59> jamiemcc: To be fair to it, strace is taking up plenty of CPU, so it's not an entirely fair test
[08:32] <mjg59> But I want to see the syscall counts
[08:32] <jamiemcc> ok
[08:33] <mjg59> (Still running)
[10:46] <mathiaz> stgraber: I'm thinking about asking the ubuntu-server mailing list to do some hardware testing of tribe-4. Do you have an idea on how the iso testing tracker could be used to track the results ?
[10:57] <Kmos> soren: ping
[10:59] <mathiaz> Kmos: soren is on vacation for at least two weeks.
[11:00] <Kmos> mathiaz: ok thx :(
[11:06] <stgraber> mathiaz: this will be easily possible once the qa.stgraber.org (maybe qa.ubuntu.com by this time) will be working and we'll be able to create sub-domain
[11:07] <stgraber> mathiaz: but I'm still starting implementing it (kernel and mozilla team being the testers), the details should be discussed at next UDS
[11:08] <stgraber> mathiaz: so, nothing now :(
[11:10] <mathiaz> stgraber: ok. Do you think we can add test case per hardware type ?
[11:10] <mathiaz> stgraber: for example 'dell poweredge 850'
[11:11] <stgraber> inside the ISO tracker nope, but in the future you will be able to request a server.qa.stgraber.org website and there it'll be possible
[11:12] <stgraber> you'll set Gutsy+1 BETA or whatever the CD is as milestone
[11:12] <stgraber> then add your Hardware where you currently see the different ISOs
[11:12] <stgraber> then add you testcase on them
[11:12] <mathiaz> stgraber: hum... I see. I'm trying to come up with a solution that would be usable now.
[11:13] <mathiaz> stgraber: we have the checkCD test case, the LAMP test case.
[11:13] <mathiaz> stgraber: is it possible to add a "DellPoweredge" test case ?
[11:13] <stgraber> mathiaz: yes
[11:14] <mathiaz> stgraber: ok. And can I do this ?
[11:14] <mathiaz> stgraber: or should I contact you ?
[11:14] <stgraber> but everyone will see it (both server team and classical contributor) and some people may wonder what the hell is that when doing testing on server iso :)
[11:15] <stgraber> mathiaz: testcase management is still something I'm the only one who can do that (have to put that directly in the DB), there is still no UI for that (yet)
[11:15] <mathiaz> stgraber: ok. What about 'HW: DellPoweredge' ?
[11:16] <mathiaz> stgraber: I don't have the list of hardware yet. Just discussing the possibilities.
[11:18] <stgraber> yes, that can be a temporary solution before you have your own tracker set (which I think would be better)
[11:20] <mathiaz> stgraber: yes. I agree with you.
[11:20] <mathiaz> stgraber: but that won't happen before sometime.
[11:21] <stgraber> Well, Mozilla team will have some kind of beta working for next Firefox update, but that'll still look like the ISO tracker (lot of ISO specifc things)
[11:22] <stgraber> What the spec for Gutsy+1 will mainly add is usage specifc modules and user profile options (btw, I'll add the server team as being interested in the new QA tracker in the Spec)
[11:22] <mathiaz> stgraber: please do. We're interested that too.
[11:25] <stgraber> mathiaz: as you don't really need much "server testing" specific thing, I'll add you to the "beta test" which could soon be ready (that's just a matter or fixing some Admin UI, user UI is already ready)
[11:28] <mathiaz> stgraber: great. I guess I'll come up with new ideas/feature proposal once I start using the tracker... :)
[11:41] <justinwray> bdmurray: May I have a word with you?
[11:47] <manchicken> Does anybody know whether or not the Ubuntu Dells (running 7.04) are running OEM with Beryl and/or compiz?
[11:47] <Amaranth> manchicken: doubtful
[11:47] <manchicken> I asked Dell, and Dell told me that they're running Ubuntu.
[11:47] <manchicken> heh
[11:48] <Amaranth> afaik they use stock 7.04
[11:48] <manchicken> I would think so, too.
[11:48] <manchicken> Amaranth: Long time no talk.  You're the guy from UDS, right?
[11:49] <Amaranth> yep
[11:49] <manchicken> Thought so :)
[11:49] <Amaranth> manchicken: do you know if it's a boy or girl yet?
[11:49] <manchicken> Not yet.
[11:49] <manchicken> We'll know in one week.
[11:49] <Amaranth> Got pretty worried about you that last night
[11:49] <manchicken> Yeah.
[11:50] <manchicken> It was not a pretty trip home.
[11:50] <LaserJock> oh?
[11:50] <manchicken> Got well home and got much better real quick.
[11:50] <LaserJock> you get sick?
[11:50] <Amaranth> I was thinking exhaustion
[11:50] <manchicken> Yeah, evidently if you get foreign bacterial infections, drinking beer is not a good idea.
[11:50] <manchicken> Amaranth: Dehydration.
[11:50] <Amaranth> Oh, even better :/
[11:50] <LaserJock> I got really sick at the end of UDS Paris
[11:50] <manchicken> Amaranth: I got a nasty water-borne bacterial infection.
[11:51] <LaserJock> took me at least 2-3weeks to get better
[11:51] <Amaranth> manchicken: That explains it, you're not supposed to drink the local water
[11:51] <LaserJock> but UDS Sevilla was good for me (other than not a lot of sleep)
[11:51] <manchicken> I didn't.
[11:51] <manchicken> I think I got it from the shower.
[11:51] <Amaranth> Could be
[11:51] <cjwatson> I think my illness at UDS-Sevilla was probably caffeine withdrawal
[11:52] <cjwatson> though it was a bit much for that, might have been a bug too
[11:52] <manchicken> What's scary is I was drinking about 2-3 liters of water a day.
[11:52] <Amaranth> That hotel worker thought you were wasted
[11:52] <manchicken> These Dells have the intel GPUs, but still run Intel wifi cards.
[11:52] <manchicken> Dude, totally.
[11:53] <manchicken> The security guy was about to arrest me until that Beryl guy's mom stopped in.
[11:53] <Amaranth> err, intel GPU + intel wireless is common
[11:53] <manchicken> stepped*
[11:53] <manchicken> Yeah, unfortunately the intel wifi is binary-blobish.
[11:53] <Amaranth> manchicken: Oh, I thought you were out when that happened :)
[11:53] <Amaranth> manchicken: iwlwifi
[11:53] <Amaranth> only needs firmware
[11:53] <manchicken> I was coming to when that happened.
[11:54] <manchicken> I wasn't out for long.
[11:54] <Riddell> infinity: could you give back strigi
[11:54] <manchicken> Let that be a lesson.  Carry a thermometer with you when you travel, and take it easy when you're sick.
[11:54] <Nafallo> Amaranth: isn't there some binary application as well?
[11:54] <manchicken> that is, if homeland security will let you onto an airplane with a thermometer.
[11:55] <manchicken> Amaranth: Yeah, and that firmware is binary-blob.
[11:55] <mjg59> Nafallo: Not for 4965
[11:55] <mjg59> Yes for 3945
[11:55] <manchicken> And the dells come with 3945
[11:55] <Amaranth> mjg59: iwl3945 requires a daemon?
[11:55] <Nafallo> mjg59: thanks. I thought I remembered something about it :-)
[11:55] <Amaranth> mjg59: I thought only ipw3945 did
[11:55] <Nafallo> manchicken: the Dells often let you choose, no?
[11:55] <mjg59> Amaranth: iwlwifi is not currently usable for 3945
[11:56] <manchicken> Nafallo: No
[11:56] <Amaranth> mjg59: I noticed :/
[11:56] <manchicken> mjg59: What is?
[11:56] <manchicken> mjg59: Please don't say ndiswrapper
[11:56] <Amaranth> Nafallo: If they did it'd be a choice between ipw3945 and broadcom
[11:57] <manchicken> I will be *very* upset if I buy a machine and have to run ndiswrapper.
[11:57] <LaserJock> does madwifi work for most Atheros chips?
[11:57] <manchicken> IIRC
[11:57] <manchicken> I've never had one of those.
[11:57] <manchicken> I've got a broadcom.
[11:57] <manchicken> Which is the scourge of mankind.
[11:57] <Nafallo> Amaranth: for D630 I could choose between 3945 and 4965...
[11:58] <mjg59> manchicken: ipw3945
[11:58] <manchicken> Nafallo: Dell is only giving out 3945, and I think System76 is doing much of the same.
[11:58] <Amaranth> Nafallo: Oh, right, broadcom is only on the Windows versions
[11:59] <Nafallo> Amaranth: only US has non-Windows versions probably :-)
[11:59] <manchicken> I ordered my dell on the 23rd of July, they still haven't shipped it or charged the money.
[11:59] <LaserJock> hmm, I suppose "Atheros Communications, Inc. Unknown device 001c (rev 01)" is not a good sign
[11:59] <Nafallo> Amaranth: at least D630 only came with Windows :-P
[12:00] <manchicken> Nafallo: I think Dell sells Ubuntu machines wherever they ship to normally.
[12:00] <Nafallo> manchicken: I ordered one 5th of July. When they delayed it until 22nd of August I called and told them I couldn't wait that long.
[12:00] <manchicken> What's the alternative?  Pay the windows tax?
[12:00] <Nafallo> manchicken: yes.
[12:00] <Nafallo> manchicken: or freedos
[12:00] <Amaranth> Dude I think I read on a blog the other day about someone ordering from Dell when the announcement was made and still not having their laptop
[12:00] <manchicken> Freedos isn't a bad idea if the hardware is decent.
[12:01] <manchicken> But I wanted a laptop, and I don't think they have any lappies with freedos
[12:01] <manchicken> Amaranth: They're saying they'll for sure have it out as of the 17th, so we'll see.
[12:01] <manchicken> I just hope they get it to be before we close on this house and move.
[12:01] <Amaranth> don't believe the date on the website
[12:01] <Amaranth> Unexpected demand and all
[12:02] <LaserJock> I'd love to know what the sales figures have been
[12:02] <bhale> hi LaserJock
[12:02] <Nafallo> Amaranth: mine was delayed cause they couldn't get the LCD in time :-)
[12:02] <LaserJock> bhale:  hi!
[12:03] <LaserJock> once Walmart starts selling Dells with Ubuntu I'll be impressed
[12:03] <LaserJock> might even try to buy one
[12:03] <LaserJock> :-)
[12:04] <manchicken> Amaranth: Well, they estimated more than three weeks out, I would hope they'd be right.
[12:04] <ajmitch> hello
[12:04] <LaserJock> hola ajmitch
[12:05] <mathiaz> hi ajmitch
[12:05] <ijuz_> i got btw. my Dell with FreeDOS
[12:05] <manchicken> ijuz_: Was it a laptop?
[12:05] <ijuz_> the Ubuntu laptop offers aren't so cheap and halfway crappy stuff
[12:06] <ijuz_> yes, a Latitude D830 n-series (that means without OS basically)
[12:06] <bhale> right, i buy latitude (business) over the home user stuff
[12:06] <stgraber> LaserJock: with only some ads in Newspapers (as they do with their weekly offers) would already impress me :)
[12:08] <manchicken> I don't know, the laptop I ordered looks awful nice on paper.
[12:08] <LaserJock> I think I'll just buy Apple from  now on
[12:09] <manchicken> They do broadcom.  No thanks.
[12:09] <ijuz_> i asked in an apple store, you can't get apple laptops without OS
[12:09] <manchicken> And ATI
[12:09] <LaserJock> ijuz_: well, I want the OS, that's partly the point
[12:09] <stgraber> I personaly bought one HP Compaq (enterprise product), which was SLED certified (but shiped with XP Pro and free upgrade to Vista business)
[12:10] <manchicken> ijuz_: And you want the ATI and the Broadcom?
[12:10] <LaserJock> ATI doesn't bother me so much, Broadcom might
[12:10] <manchicken> I've been on Broadcom and ATI for 2 years now.  Never again.
[12:10] <manchicken> And I'll never buy another HP again.
[12:10] <ijuz_> manchicken: no thanks... i have that in my latitude D810 ;)
[12:10] <stgraber> so I have both XP (school use), Vista (support for people not running Linux yet) and SLED certified (any Linux should run just fine)
[12:11] <LaserJock> I've got an HP desktop from walmart, it was a piece of junk, but from Walmart, what do you expect
[12:11] <manchicken> stgraber: That's fine, but I don't want to run windows.
[12:11] <manchicken> stgraber: I have no interest in buying them.
[12:12] <manchicken> stgraber: I just want a standard Ubuntu OEM with hardware that works well with free software.  Dell came the closest to that.
[12:13] <manchicken> LaserJock: The HP is more of an indicator than Walmart.
[12:13] <LaserJock> manchicken: what about System76?
[12:13] <Amaranth> manchicken: what wifi works without firmware?
[12:13] <Amaranth> my HP works great
[12:13] <manchicken> LaserJock: They're more pricy than Dell, but they're okay.